1. Introduction
As academic scholarship proliferates under the umbrella of International Disaster Law (IDL), it naturally prompts rumination as to what the phrase is being used to mean. It is commonly asserted, and correctly so, that there is no discrete branch of IDL.Footnote 1 The British Red Cross, for instance, insists on referring to ‘international disaster laws’.Footnote 2 However, this proposition is not correct because the instruments of IDL, as yet, lack the normativity, depth or coherence of other so-called ‘branches’.Footnote 3 It is also correct because no ‘branch’ has a ‘legally meaningful existence’ in international law.Footnote 4 No branch is truly ‘self-contained’;Footnote 5 the system of international law operates as an interconnected body of norms, with more general and foundational norms pervading each area of particularized regulation on subject matters such as disasters. As the International Law Commission (ILC) has recognized, a so-called ‘branch’ ‘can receive … legally binding force … only by reference to (valid and binding) rules or principles outside it’.Footnote 6
Necessarily, then, the birth and trajectory of IDL has mirrored this normative interdependence. IDL codification efforts since 2000 have yielded several soft law instruments governing state responses and co-ordination at every point along the ‘disaster cycle’,Footnote 7 from prevention and ‘disaster risk reduction’ (DRR) to response, recovery, and compensation. Chief among these instruments are the 2015 Sendai Framework,Footnote 8 and the ILC’s 2016 Draft Articles on the Protection of Persons in the Event of Disasters,Footnote 9 the latter of which the ILC recommended to the UN General Assembly as the basis of a future treaty in the area.Footnote 10 Though heterogeneous in their content and scope, these instruments all subscribe to what can be labelled an ‘adaptive’ model of disaster regulation, in terms of their relationship to the greater system of international law: they seek to develop, adapt, and particularize the application of norms from other, more established subfields to disaster situations – and where necessary, to crystallize new norms. This has served the overriding objective to produce a more holistic body of law and fill the ‘yawning gap’ once suggested to exist at the heart of IDL.Footnote 11
To this end, various norms and concepts from international human rights law (IHRL),Footnote 12 humanitarian law (IHL),Footnote 13 and trade law,Footnote 14 among others,Footnote 15 have been adapted and particularized to varying extents. For instance, Article 5 of the ILC Draft Articles confirms that persons affected by disasters are entitled to ‘respect for and protection of their human rights in accordance with international law’, with the reference to ‘human rights’ being to ‘the whole of international human rights law’.Footnote 16 Similarly, the Sendai Framework includes ‘promoting and protecting all human rights’ among its guiding principles.Footnote 17 Through this iterative process of particularization has emerged the ‘overarching umbrella’ that characterizes IDL’s present structure.Footnote 18
Even so, these instruments do not capture the full extent of the relationship that other subfields of international law may have with disasters – the true scope of IDL, as it were. This ‘adaptive’ model to disaster regulation prevalent in all recent IDL initiatives exists alongside a ‘reactive’ model. The latter model encompasses those means, mechanisms, and techniques under international law by which other subfields treat disasters as exogenous events which justify or necessitate deviations from the ordinary application of legal norms on their subjects. The difference between the models is that, whereas norms and instruments embodying an adaptive model seek to reconcile norms with disaster situations, those embodying a reactive model seek to reconcile disasters with norms. In other words, adaptive regulation develops norms that do apply in disaster situations, while reactive regulation serves to identify those that do not. Together, these models comprise the current corpus of IDL, though the reactive model remains under-examined.
The focus of this article is to present and critique this reactive model for disaster regulation. Its contribution to the existing IDL literature is that it offers a single theoretical framework to bring together certain interfaces between international law and disasters (those embodying this reactive model) which each fall within the conceptual scope of IDL and may have been individually scrutinized as such,Footnote 19 but which have not been theorized from a systemic, cross-sectional perspective. This systemic view allows one to trace, historicize, and explain the evolution in disaster regulation toward the adaptive model characteristic of IDL today. In so doing, it addresses the ‘core [epistemological] challenge’ to ‘achieve a more comprehensive account of IDL’.Footnote 20
This article has four subsequent sections. Section 2 introduces the reactive model of disaster regulation, by which disasters are reconciled with a state’s existing obligations under international law. It identifies two regulatory techniques that embody this reactive approach: ‘disapplication’ and ‘exculpation’. Section 3 then identifies various doctrines and mechanisms through which these techniques have been operationalized, both within general international law and ‘special treaty-regimes’.Footnote 21 It also draws on previous scholarship to identify doctrinal impediments within these mechanisms that may problematize their application to disasters.
On this basis, Section 4 then considers whether the techniques of disapplication and exculpation, and the reactive model overall, are conceptually appropriate for disaster regulation. It argues that certain disapplication and exculpation mechanisms are maladapted to regulate disasters because they rely on an anachronistic, long-delegitimized view of disasters as rare and episodic events beyond human control. This puts such mechanisms at odds with the core objectives of modern IDL, and the sociological and scientific premises on which they are based, which emphasize the need for human co-ordination and preventive regulation to counter the rising regularity and severity of disasters. Conversely, those mechanisms which can accommodate this modern view of disasters and IDL’s corresponding imperatives remain more readily applicable at a doctrinal level. In sum, this both exposes conceptual frailties within the reactive model of disaster regulation and explains IDL’s inclination toward adaptive regulation. Section 5 discusses the implications of these findings for future research and policymaking in the field.
For the purposes of this article, ‘disaster’ is understood in line with the ILC Draft ArticlesFootnote 22 to mean ‘a calamitous event or series of events resulting in widespread loss of life, great human suffering and distress, mass displacement, or large-scale material or environmental damage, thereby seriously disrupting the functioning of society’. It includes both sudden and slow-onset events, health-related events (including the COVID-19 pandemic), and both ‘natural’ and ‘human-induced’ events, given the ‘porous border’ between the two recognized in modern sociology.Footnote 23 However, it excludes armed conflicts, in which rules of international humanitarian law primarily apply,Footnote 24 and ‘economic’ events such as financial crises per se, mirroring the same distinction drawn across IDL instruments.Footnote 25 Nonetheless, this article does later analyse some of the investor-state dispute settlement (ISDS) jurisprudence arising out of such economic events (especially the 2001–2002 Argentine financial crisis) insofar as it implicated the same norms and mechanisms that are potentially invocable in disaster situations.
2. The reactive model of disaster regulation and its constituent techniques
This article introduces the notion of a ‘reactive’ model of disaster regulation in international law. As explained above, whereas the adaptive model is concerned with norms that do apply in relation to disasters and regulatory responses thereto, the reactive model concerns doctrines and mechanisms that govern whether and to what extent those norms do not apply. While these doctrines are strewn across different subfields, bear different names, and apply in different contexts, it is possible to distil the model into two distinct groups of doctrines and mechanisms, based on their functionally similar approaches to reconciling norm with disaster. This article describes each group as representing a ‘technique’ of disaster regulation, though use of this term does not imply that they were ever deliberately conceived as such in the development of international law.
The first technique involves ‘disapplication’ of an ordinarily applicable norm in relation to a disaster or a state’s disaster-related measures. This refers to situations where a norm of international law which would otherwise bind a state is deemed not to apply, either unilaterally by the state concerned or during adjudication. The second involves ‘exculpation’ of a state from the consequences of acting incompatibly with a norm in circumstances of disaster. This refers to situations where a breach of a primary norm is prima facie established, but a state is excused from the ordinary consequences of such breach. Though this distinction is narrow and sometimes collapses in practice,Footnote 26 the two categories are mutually exclusive: a norm that is disapplied (evidently) does not apply, meaning that no consequences arise from acting incompatibly with it; and only if a norm does apply (and thus has not been disapplied) can a state be exculpated from the consequences of breaching it.
Such a distinction between ‘disapplication’ and ‘exculpation’ has been commonly made, though not in these terms and not in specific relation to disasters, in various contexts. For instance, it has been articulated as a difference between the characterization of clauses in investment protection treaties as ‘exemptions’ or ‘exceptions’;Footnote 27 between ‘scope limitation devices’ and ‘affirmative defences’ for the purposes of the burden of proof in international adjudication;Footnote 28 and between ‘justification’ and ‘excuse’ in the context of how defences in the law of state responsibility operate theoretically.Footnote 29 Disapplication mechanisms are also said to operate as ‘primary’ norms (i.e., those ‘defining conduct or obligations’Footnote 30), whereas exculpation mechanisms are ‘secondary’ norms of state responsibility (i.e., ‘concerned with determining the consequences of failure to fulfil obligations established by the primary rules’Footnote 31).Footnote 32 These particular terms have thus been selected for their generality and descriptive convenience, though there are also legally significant differences between them, e.g., in respect of burdens of proof and interpretive approaches during adjudication.Footnote 33
On this basis, the following section examines examples and modalities of the techniques of disapplication (Section 3.1) and exculpation (Section 3.2) across the breadth of international law, drawing on literature which critiques the applicability of each modality to modern-day disasters.
3. Disapplication and exculpation mechanisms across the breadth of international law
3.1 Disapplication mechanisms
States are ordinarily bound to perform their obligations under international law in good faith (pacta sunt servanda).Footnote 34 However, this principle – this ‘staccato statement’Footnote 35 – is not absolute and indefeasible. Certain mechanisms exist by which a state can, owing to extraordinary circumstances, ‘disapply’ a primary norm with the effect that it is not bound by the norm for a particular period.
‘Disapplication’ is given effect through mechanisms invocable under general international law (Section 3.1.1) and in special treaty-regimes (Section 3.1.2); or otherwise in the course of an international court or tribunal’s process of interpreting a primary norm (Section 3.1.3).
3.1.1 Under general international law
Unlike customary international law, which by its nature is based on a consistent pattern of conduct, treaties are ‘in permanent tension to the passing of time and changing circumstances’,Footnote 36 and are thus susceptible to becoming ‘dead letter’ if and when circumstances radically change, e.g., in times of disaster.Footnote 37 In order to imbue treaties with some capacity to accommodate change without becoming obsolete, the Vienna Convention on the Law of Treaties (VCLT) codifies default rules, recognized as custom,Footnote 38 allowing for suspension and termination of treaties.
In the first instance, a state can suspend,Footnote 39 denounce or withdraw from,Footnote 40 a treaty at any time, if permitted by the treatyFootnote 41 or with consent of all parties. Furthermore, a state has the unilateral right, recognized in custom,Footnote 42 to suspend or terminate a treaty in extraordinary circumstances – potentially including disasters – which give rise to ‘supervening impossibility of performance’Footnote 43 or a ‘fundamental change of circumstances’.Footnote 44 Upon suspension, the state is released from its obligations under the treaty for the relevant period.Footnote 45
These provisions are rarely invoked to disapply treaty norms, either generally or during disasters.Footnote 46 This may be a consequence of their restrictive wording and strict substantive criteria for enlivenment,Footnote 47 borne from a cautiousness not to allow these residual disapplication mechanisms to become ‘talisman[s] for revising treaties’.Footnote 48 Instead, parties are encouraged to provide expressly in advance for denunciation or withdrawal, through ‘exit clauses’ and other flexibilities.Footnote 49
However, these mechanisms are also poorly adapted to the reality of disasters and their impacts. For one, supervening impossibility of performance under Article 61 requires the ‘permanent disappearance or destruction of an object indispensable for the execution of the treaty’.Footnote 50 The ‘object’ requirement, which is viewed to encompass physical and (perhaps) juridical objects,Footnote 51 does not extend to intangible institutional (e.g., financial, diplomatic, labour) resources that are typically affected in a disaster, especially one of a non-environmental character (e.g., a pandemic). In particular, the ICJ has expressly ruled out ‘serious financial difficulties’ from the scope of the condition.Footnote 52 Similarly, the requirement of ‘permanency’ may exclude the destruction of institutions or infrastructure which could be restored, albeit over time and at great cost, as occurs in many disasters.
The way these conditions are framed risks inconsistent and perverse results when applied to disasters. For instance, the complete and permanent drying up of an international watercourse because of drought would satisfy the ‘object’ requirement in respect of any treaty governing that watercourse.Footnote 53 Such treaties are easily reconcilable with Article 61 because they gravitate around one physical ‘object’ capable of destruction. Conversely, in recent years, Haiti has endured a 2010 earthquake that caused US$ 8 billion of damage to its infrastructure and economy (equivalent to 70% of its Gross Domestic Product at the time),Footnote 54 a 2016 hurricane causing damage equal to 32 per cent of its GDP,Footnote 55 and consistent periods of drought that devastated its food supply.Footnote 56 But none of these disasters, it seems, would have entitled it to suspend investment treaties with Germany, France, and the United Kingdom on Article 61 grounds. While each impaired its financial and institutional ability to afford the requisite level of treatment to foreign investors, the ‘objects’ indispensable to execution of those treaties – if conceived rather broadly and circularly as a functioning state apparatus with a legal, economic and regulatory environment capable of investment protection – remained in existence or only temporarily ‘destroyed’. Perversely, the scope of the ‘object’ requirement suggests that only if the Haitian state apparatus had reached total collapse would its right to suspend investment treaties under Article 61 have enlivened.
Similarly, a ‘fundamental’ change under Article 62 requires a ‘radical transformation’Footnote 57 of what must be performed under the treaty into ‘something essentially different’.Footnote 58 Again, this would likely not capture an instance where a disaster (e.g., the 2010 Haitian earthquake) adversely affects a state’s financial and technical capacity to perform an obligation for a particular period, but not the basic nature of the obligation itself. Furthermore, the requirement that the circumstances be ‘not foreseen by the parties’ risks becoming an insurmountable condition in modern times. As discussed in Section 4, it is the change in circumstances, not the disaster causing the change, which must be ‘foreseen’: in this way, the veritable mountain of climate change research indicating the increased likelihood and regularity of disasters renders it difficult for a state to argue that it has not foreseen such a change.
Accordingly, disapplication mechanisms under the VCLT are blunt instruments calibrated in order to prioritize treaty stability. This means they are unlikely to be successfully engaged in most disaster contexts.Footnote 59
3.1.2 In special treaty-regimes
Apart from the law of treaties, there are also disapplication mechanisms, constructed differently but identical in effect, in IHRL (Section 3.1.2.1), and investment agreements (Section 3.1.2.2).
3.1.2.1 Derogation in IHRL
Certain international and regional human rights treaties allow a state to ‘derogate’ from (i.e., suspend) specified obligations under the treaty.Footnote 60 This qualifies as a ‘disapplication’ mechanism because it results in the norm not applying for the relevant period, rather than a state avoiding responsibility for acting incompatibly with it.Footnote 61 A state’s right to derogate represents a corollary of sovereignty, through which states are afforded a ‘safety valve’ to tailor certain human rights standards in an extreme situation of emergency.Footnote 62 Indeed, the 2016 ILC Draft Articles preserve a state’s right to derogate, with the commentary to Draft Article 5 noting that the provision ‘incorporates both the rights and limitations that exist in the sphere of international human rights law… [and] contemplates an affected [s]tate’s right of suspension or derogation where recognised under existing international agreements’.Footnote 63
Derogation clauses in international and regional treaties are formulated similarly.Footnote 64 The clause in the International Covenant on Civil and Political Rights (ICCPR), for instance, may be invoked ‘in times of public emergency which threaten the life of the nation’,Footnote 65 a threshold within which severe natural disasters have been regarded to fall.Footnote 66 The power to derogate is subject to certain limitations, including a carve-out of certain non-derogable rights, a proportionality requirement that derogation must go no further than ‘strictly required by the exigencies of the situation’ in terms of duration, geographical coverage, material scope, and measures adopted,Footnote 67 and requirements of initial proclamation and ongoing reporting.Footnote 68 In this way, they are ‘inherently temporary’ and ‘follow the very concept of a state of emergency’.Footnote 69
Historically, derogation has rarely formed part of a state’s disaster response. Sommario identified just five states up to 2018 that invoked ICCPR Article 4 in relation to disasters.Footnote 70 However, this must be qualified in two respects. First, there has been a relative spike in state practice of derogation during the COVID-19 pandemic. Since the WHO’s initial declaration of a pandemic on 11 March 2020,Footnote 71 32 states have derogated at some point from certain rights in the ICCPR, ECHR, and/or ACHRFootnote 72 – typically the freedoms of movement, assembly, and association – in order to implement quarantine and social distancing measures.Footnote 73 But even then, this represents a mere fraction of those states that have implemented measures requiring derogation without undergoing the mandated notification process.Footnote 74
Secondly, this historical scarcity is not necessarily explained by any doctrinal disconnect between derogation clauses and the reality of disasters. To the contrary, notwithstanding more general critiques of the ‘problematic design and operation of derogation clauses’,Footnote 75 derogation clauses are well-adapted to capture disasters in principle, given their focus on the ‘magnitude and effects of the emergency, irrespective of its nature or origin’.Footnote 76 Indeed, the requisite threshold even favours vulnerable states who may find it easier to claim the ‘life of the nation’ has been threatened by a disaster.Footnote 77
Rather, other compelling explanations exist for this scarcity. For one, political scientists have attributed it to a leniency among states toward failures to derogate, and a corresponding presumption that derogations are done in good faith and not as a ‘veil for gross abuses of human rights’.Footnote 78 Evidencing this claim, no human rights treaty body has ever examined the legality of a derogation made in disaster response.Footnote 79 Secondly, states can rely during disasters on in-built limitations within particular human rights provisions, which do not require official proclamation, instead of their right of derogation. Given the paucity of derogations during the COVID-19 pandemic relative to the ubiquity of quarantine and social distancing arrangements, most states appear to rely on the ‘public health’ exception built into the freedoms of movement, assembly, and association.Footnote 80 Indeed, Sommario argues that because of the flexibility provided by these limitations, formal derogation is rarely necessary in disaster response.Footnote 81
In sum, despite the relatively frequent state practice of derogation in general, this avenue for disapplication has rarely been employed in relation to disasters, though prevailing academic sentiment does not view this as a consequence of doctrinal unsuitability.
3.1.2.2 ‘Non-precluded measures’ clauses in investment agreements
Disapplication mechanisms also feature in international investment law, though unlike derogation under IHRL, they typically operate in an adjudication context, and retrospectively effect a disapplication. Certain international investment agreements (IIAs) contain ‘non-precluded measures’ (NPM) or ‘carve-out’ clauses,Footnote 82 which permit a state to ‘lawfully take action directed at a particular regulatory objective … that might otherwise be inconsistent with its substantive treaty obligations toward foreign investors’.Footnote 83 A well-litigated example is Article XI of the US-Argentina Bilateral Investment Treaty, which provides that the treaty ‘shall not preclude the application by either Party of measures necessary for the maintenance of public order … or the Protection of its own essential security interests’.Footnote 84
There is disagreement over whether NPM clauses operate by limiting the scope of substantive obligations in a treaty, such that measures falling within the clause are not regulated by the treaty and are beyond a tribunal’s jurisdiction (‘scope limitation’), or by providing an affirmative defence to justify conduct that would otherwise be prohibited by the treaty.Footnote 85 Generally, this depends on the ‘specific wording [of the treaty] or on an explicit formulation’.Footnote 86 However, whereas most jurisprudence on Article XI above has treated it as an affirmative defence,Footnote 87 this article follows the strident criticism of this interpretation, which argues in favour of treating NPM clauses (including Article XI) as disapplication mechanisms subject to specific language to the contrary. This is a more persuasive account of the scope of NPM clauses, which typically neither refer to, nor purport to displace or overlap with, customary defences.Footnote 88
A conceptual difference from derogation and VCLT grounds for suspension is that disapplication under NPM clauses arises not out of a disaster itself, but out of measures taken by a state in relation to it. Accordingly, the doctrinal suitability of NPM clauses for disaster regulation depends on whether and which disaster-related measures are captured. The only salient examination of this issue so far occurred in the ISDS proceedings of Devas v. India and Deutsche Telekom v. India. Footnote 89 These proceedings arose out of the Indian government’s decision in 2011 to cancel an agreement for the lease of two satellites. India argued that the decision served to protect an ‘essential security interest’ for the purposes of the NPM clauses at issue. In particular, it argued that it made the decision in order to preserve the band of electromagnetic spectrum in which the satellites functioned for military operations, but also for other ‘societal needs’ including services for ‘disaster management’, ‘emergency communication’ and ‘dissemination of disaster warnings’.Footnote 90 However, both tribunals dismissed the portion of India’s justification based on disaster management considerations, taking the view that ‘“societal needs” such as … disaster management’ could not be understood as essential security interests ‘without distorting the natural meaning of [that] term’.Footnote 91 On that basis, it appears that, at least on the view of two ISDS tribunals, DRR measures may lack the criticality and/or requisite nexus to state security to enliven an NPM clause which refers to the ‘protection of essential security interests’.
This reasoning reveals a miscalibration between the scope of NPM clauses and IDL’s core imperative of prevention. Indeed, the ILC Draft Articles specifically identify preventive measures of the kind rejected in Devas and Deutsche Telekom (e.g., early warning systems, monitoring mechanisms, community-level preparedness and education, and contingency planning) as critical components of a state’s DRR policy.Footnote 92 Nonetheless, future ISDS tribunals with a greater sensitivity to IDL and its concerns may regard DRR measures as serving ‘essential security interests’, or at least as falling within the scope of a more expansively formulated clause which extends (for instance) to measures ‘necessary to protect human, animal or plant life or health’.Footnote 93 But even once this hurdle has been overcome, a tribunal may reject reliance on the clause for want of proportionality or necessity.Footnote 94
3.1.3 Flexible interpretation of primary norms
Finally, a primary norm can be disapplied in times of disaster through a process of flexible interpretation of that norm by an international court or tribunal.
The key difference between this avenue for disapplication and derogation under IHRL is that the former cannot be invoked unilaterally or contemporaneously, whereas the latter can. Instead, flexible interpretation is undertaken during adjudication by an international court or tribunal. Moreover, the difference between this avenue and treaty NPM clauses in investment agreements is that the disapplication here occurs through interpretation of the primary norm, whereas the latter involves a disapplication of the primary norm effected by a separate and express carve-out provision.
Flexible interpretation may not be possible where the material, temporal, personal, and geographical scope of a norm is manifest. But where a norm bears some degree of generality, then a court may employ various techniques in order to qualify or dilute its scope, with a view to honouring a state’s right to regulate in situations of disaster. As Viñuales notes:
… [H]uman rights, investment standards and trade disciplines are often interpreted in a manner that introduces reasonable restrictions of the scope of a provision to allow for legitimate governmental action.Footnote 95
One technique relevant generally in international adjudication is a (peculiar) application of the doctrine of lex specialis. The doctrine has at times been adopted not in order to resolve a conflict of norms, but qua ‘articulation device’Footnote 96 – i.e., use of a specific norm to tailor the application of a more general norm to a unique situation of disaster.Footnote 97 Used in this manner, the doctrine is capable of qualifying, diminishing, nullifying, or perhaps even enhancing the application of more general norms in time of disaster.Footnote 98
However, the dearth of specific and binding norms governing disasters renders the employment of such a technique a remote prospect presently – at least, until the conversion of the ILC Draft Articles into treaty form. In that eventuality, it may be that, for instance, the specific DRR obligation in Article 9 of the ILC Draft Articles is read as a component of a state’s duty to take reasonable measures to prevent transboundary environmental harm. More generally, too, IDL norms requiring state action in DRR or disaster response may produce ‘normative conflicts’ with norms in other subfields constraining such action (e.g., investment protections or non-discriminatory trade obligations) and necessitate the use of ordinary conflicts rules, potentially resulting in disapplication of the other norm.Footnote 99
There are also other techniques specific to special treaty-regimes. For one, the ‘margin of appreciation’ doctrine, developed in European human rights jurisprudence and subsequently adopted by other human rights bodies,Footnote 100 is an approach for interpreting rights which defers to some extent to the factual conclusions drawn by national authorities as to the necessity of an impugned restriction because of their ‘direct and continuous contact with the vital forces of their countries’.Footnote 101
Moreover, the fair and equitable treatment standard in investment agreements, especially its ‘legitimate expectations’ component, is often interpreted in light of a state’s right to regulate.Footnote 102 Similarly, the statutory guarantee against uncompensated expropriations is qualified by the ‘police powers’ doctrine, which precludes certain regulatory acts undertaken in aid of legitimate public objectives from being characterized as ‘expropriations’.Footnote 103
Adaptive IDL instruments, once crystallized into hard law or perhaps even otherwise,Footnote 104 may offer a basis to invoke these disapplication doctrines. For instance, implementation of DRR measures enumerated in the ILC’s Commentary to Article 9 – e.g., ‘land-use controls, construction standards, ecosystems management, drainage systems …’ – may qualify as legitimate regulation in the public interest and enliven the police powers doctrine.
3.2 Exculpation mechanisms
The above section has explained that various disapplication mechanisms exist –though some are seldom engaged – by which ordinary norms can be deemed not to apply either during a disaster or to certain disaster-related measures. But if a norm does continue to apply, there are also various mechanisms by which a state can exculpate themselves from the consequences of acting incompatibly with that norm in relation to a disaster. These may arise in the form of customary defences (Section 3.2.1) and treaty exceptions (Section 3.2.2), both of which presuppose some inconsistency with a primary norm.
To be sure, there has been some conflation of the two techniques in practice, especially in the context of NPM clauses whose manner of operation can either be viewed as scope limitation device or as affirmative defence.Footnote 105 Indeed, some exculpation mechanisms were even designed as alternatives to disapplication mechanisms; the ‘legalization’ of defences such as force majeure and necessity was intended to accommodate subsequent events and changes without endangering the stability of an entire treaty by necessitating suspension.Footnote 106
3.2.1 Under general international law
Under general international law, there are six ‘circumstances precluding wrongfulness’ codified in the ILC Articles on State Responsibility, of which two – force majeure (Section 3.2.1.1) and necessity (Section 3.2.1.2) – are especially relevant in time of disaster. Despite some academic disagreement as to whether these mechanisms operate by rendering impugned conduct lawful (as ‘justification’) or by excusing unlawful conduct (as ‘excuse’),Footnote 107 the dominant state practice is to invoke these circumstances as affirmative defences in adjudication, after breach of a primary norm has been established.Footnote 108 The term ‘exculpation’ has been deliberately chosen as a descriptor without prejudice to these competing characterizations of ‘justification’ and ‘excuse’.
3.2.1.1 Force majeure
The defence of force majeure, embodied in Article 23 of the ILC Articles on State Responsibility,Footnote 109 and reflecting both custom and a general principle of law,Footnote 110 is available to states that have failed to comply with an international obligation in the event of an ‘irresistible force or unforeseen event’ beyond their control, which renders performance of the obligation ‘materially impossible in the circumstances’.Footnote 111 It is suggested to have ‘deep roots’ in human thinking about morality and responsibility,Footnote 112 and represents a quintessential expression of the (long debunked) belief that disasters are events beyond human control. Reflecting such a view, Argentine jurist Carlos Calvo’s definition of force majeure was as ‘an accident that the vigilance and work of man cannot prevent or impede’.Footnote 113
Though largely invoked in the past by foreign investors in times of political insecurity or war,Footnote 114 the requisite ‘unforeseen event’ can, in theory, either be natural or anthropogenic.Footnote 115 A rare invocation in response to disasters occurred in 1887, when it was successfully pleaded by the Venezuelan government before a French-Venezuelan Mixed Claims Commission to excuse its responsibility for property damage, and a resulting suspension of operations, suffered by a French construction contractor because of floods, fires and earthquakes, among other things.Footnote 116
The doctrine has at times been criticized for providing a carte blanche to states to breach their international obligations – a ‘lawless space where everything goes’.Footnote 117 Rougier suggested as early as 1903, albeit in the context of civil wars, that the logic of absolute non-responsibility behind force majeure could not be accepted, at least not insofar as it would permit a state to excuse its conduct for an indefinite period after an unforeseen event has occurred.Footnote 118 Applying this view to disasters, a state may justifiably evade responsibility for damage to an investment caused directly by an earthquake, but not automatically in respect of every subsequent measure it enacts as part of disaster response.
In this way, the COVID-19 pandemic is a useful illustration of the extent of force majeure’s relevance to modern disasters. As Paddeu and Jephcott have observed,Footnote 119 while quarantine and social distancing measures may well adversely affect foreign investments, it may be difficult for states to rely on force majeure to avoid the consequences of failing to perform international obligations owed in respect of such investments. Two of its conditions are particularly vexing. First, states affected by the virus after it was declared a pandemic could hardly rely on the proposition that the triggering event was ‘an unforeseen event’. Nor would it qualify as an ‘irresistible force’, in the sense of an ‘element of constraint which the state was unable to avoid or oppose by its own means’,Footnote 120 if a court or tribunal took the view that the virus could have been prevented by more timely border closures or prudent social distancing measures. Conversely, given mathematical models which suggest that measures such as contact tracing are ineffective in controlling the virus in light of its asymptomatic transmission, the opposite conclusion may be just as forthcoming.Footnote 121 Secondly, the requirement of ‘material impossibility’ of performance, if understood to require absolute impossibility,Footnote 122 may not be easily satisfied if one takes the position that states could, albeit with considerable cost to human life, continue to operate unrestricted, and that these measures are technically voluntary.
Accordingly, force majeure appears a blunt and unwieldy instrument for disaster regulation. As Section 4 will argue, this doctrinal unsuitability is a result of its reliance on an anachronistic sociological view of disasters.
3.2.1.2 Necessity
Necessity is also available as a customary defence where an act not in conformity with an international obligation is the ‘only way for [a] State to safeguard an essential interest against a grave and imminent peril’.Footnote 123 It is subject to the conditions that it must not ‘seriously impair an essential interest’ of the state(s) to which the obligation is owed or the international community as a whole,Footnote 124 and that it cannot be invoked if the obligation itself excludes its availability, or if the state has contributed to the situation of necessity.Footnote 125
Again, there is some inconsistency in the jurisprudence as to how the customary plea of necessity interacts with treaty NPM clauses.Footnote 126 Views diverge as to whether such clauses import,Footnote 127 displace as lex specialis,Footnote 128 or apply independently of,Footnote 129 the customary standard.Footnote 130 This may depend on one’s approach to the anterior question of how the treaty clause operates; for instance, Henckels and Kurtz compellingly argue that treaty clauses, when understood properly as scope limitation clauses operating at the jurisdiction stage, are applied independently of the customary defence of necessity.Footnote 131
Disasters are potentially captured by the defence, though two factors render this contestable in any disaster scenario. For one, there is significant variability in its application, across ISDS tribunals at least; the 2001–2002 Argentine financial crisis (though not a ‘disaster’ per se) was at once found to constitute,Footnote 132 and not constitute,Footnote 133 a ‘grave and imminent peril’, and to threaten,Footnote 134 and not threaten,Footnote 135 an ‘essential interest of the State’.Footnote 136 Moreover, the stringency of its conditions – particularly four – is problematic. First, the ‘essential interest’ condition prompts the same concerns raised earlier in relation to NPM clauses. Based on the reasoning in Devas and Deutsche Telekom, the requisite essentiality of the interest being protected may cover measures adopted in the throes of a severe disaster, the impacts of which are already manifest, but not DRR measures adopted in advance, undermining the emphasis within IDL scholarship on prevention.
Secondly, the requirement of ‘grave and imminent peril’ sets a high threshold for disasters to surmount. Paddeu and Jephcott suggest that the COVID-19 pandemic, given its significant duration, infection and mortality rate, and cost to human life, amounts to a ‘grave and imminent peril’.Footnote 137 However, other disasters with localized impacts (e.g., mudslides affecting one town) may not be sufficiently grave to imperil an interest of the state as a whole; and long-term, climate change-induced disaster impacts such as desertification and rising sea levels may not pose a sufficiently ‘imminent’ threat at any one point in time, especially in light of the peculiar difficulties in producing scientific evidence to substantiate imminence in cases of environmental harm.Footnote 138 The requirement of imminence would also seem to preclude a plea of necessity in respect of preventive measures of the kind encouraged within IDL.
A third, potentially insuperable hurdle is the requirement that an impugned measure be the ‘only way’ to safeguard a state’s essential interest from a disaster. In relation to the COVID-19 pandemic, it may be extremely difficult to establish to a court or tribunal’s satisfaction that measures undertaken by a state were the ‘only way’ to protect human life, even with the requisite deference given to the state in this assessment. Indeed, the sheer variance between states’ quarantine policies militates against the view that there is only one way to protect human life from the virus. Further, what constitutes the ‘only way’ depends on the progress of the pandemic in any given state; border measures are not necessary once the virus has already spread within the local community. Indeed, the ‘only way’ requirement has been the decisive hurdle in defeating most international claims grounded on necessity.Footnote 139
Finally, the issue of contribution may preclude reliance, whether this requires an element of fault,Footnote 140 or mere omissions.Footnote 141 As Waibel remarked about the Argentine economic crisis, ‘[n]o ICSID tribunal will ever be able to disentangle the exogenous and endogenous causes of such crises’.Footnote 142 Again, this point is taken further in Section 4, which argues that the requirement of non-contribution is even more difficult to satisfy in light of IDL’s emphasis on the enhanced DRR responsibilities of states. Nonetheless, it is apparent more generally that the doctrine of necessity is poorly calibrated to the on-the-ground exigencies of disasters, even without regard to the modern sociology of disasters.
3.2.2 In special treaty-regimes
The technique of exculpation has not been employed only in the form of customary defences. One important specialized exculpation mechanism exists in international trade law, in Article XX of the GATT.Footnote 143 (NPM clauses in investment agreements constitute another, if interpreted as defences.Footnote 144)
Generally classified as affirmative defences,Footnote 145 the sub-paragraphs of Article XX justify the breach by a WTO member state of a primary norm of trade law if its impugned measure was ‘necessary’ for certain regulatory objectives, including the protection of public moralsFootnote 146 or human, animal or plant life or health,Footnote 147 the conservation of exhaustible natural resources,Footnote 148 or compliance with other laws and regulations.Footnote 149 Further, the chapeau of Article XX independently requires that the measure must not constitute arbitrary or unjustified discrimination between states where the same conditions prevail.
Article XX is potentially invocable in various disaster contexts, despite a dearth of jurisprudence in this regard. For instance, it appears that various states’ COVID-19 response measures, including export restrictions on personal protective equipment and medication, may violate WTO norms (e.g., Article XI of the GATT, which disallows quantitative restrictions on exports); but that Article XX(b) is applicable because the restrictions are necessary to protect human life or health.Footnote 150 Moreover, there may be circumstances in which trade restrictions must be imposed as part of compliance with a binding and sufficiently specific disaster governance framework under national law, of the kind encouraged by IDL instruments such as the IFRC’s Disaster Law Checklist and IDRL Guidelines,Footnote 151 so as to engage Article XX(d).
At the same time, this is conditional on satisfaction of the chapeau requirement of non-discrimination, and on a proportionality analysis. Indeed, the existence of an alternate, more WTO-consistent measure to achieve the same objective raises the likelihood of a measure not being justified under Article XX(b).Footnote 152 This imposes a significant regulatory burden on a state to factor WTO compliance considerations into disaster policy.
4. The modern sociology of disasters and its implications
Section 2 introduced a ‘reactive model’ of disaster regulation, encompassing the means and mechanisms by which disasters are reconciled with a state’s existing obligations under international law. This model consists of the two techniques of ‘disapplication’ and ‘exculpation’ and their constituent doctrines and mechanisms. Section 3 then outlined these doctrines and mechanisms, using both existing literature and novel analysis to support the analytic claim that certain of them – particularly the VCLT grounds for treaty suspension, and customary defences – are maladapted to capture the full spectrum of disasters.
Crucially, a further analytic claim follows: these instances of doctrinal inadequacy arise because the applicability of these doctrines to disasters relies on an outdated view of disasters as rare and episodic events beyond human control. An evolution since the mid-twentieth century in sociological and scientific thinking about disasters, then, has produced a corresponding shift in the trajectory and present orientation of IDL as a discipline, culminating in the adaptive model characteristic of IDL today. This evolution is observable in three respects.
First, it has long been understood that anthropogenic climate change continues to increase the likelihood, frequency and severity of natural hazards and extreme weather events.Footnote 153 The ILC Draft Articles are among IDL’s adaptive instruments to cite such ‘frequency and severity’ as a specific basis for regulation.Footnote 154 Conversely, the techniques of disapplication and exculpation depend to some extent on the rarity of disasters, and so their increased regularity strains them in three senses. First, it necessitates such regular recourse to these mechanisms as to significantly undermine the effectiveness and normative character of the primary norms being breached. Secondly, it also renders it impractical for a state, even if neither impecunious nor resource-strapped initially, to rely alone on disapplying norms or exculpating themselves from breaches during or after a disaster, given that costs and resource shortages will accumulate and compound over the course of a string of disasters.Footnote 155 But thirdly, as the norms within existing and future IDL instruments crystallize into binding law, then reliance on the reactive model to seek disapplication of or exculpation from a growing body of norms will become ever more counter-intuitive. The escalating ubiquity and variety of climate change-generated disaster risk, therefore, begets an adaptive regulatory approach dealing with every stage of the ‘disaster cycle’Footnote 156 – prevention and mitigation, response, recovery, and compensation – of the kind embodied in the Sendai Framework and IDL’s basic imperatives.Footnote 157
Secondly, it is well-understood – and has been for several decades in the social sciences,Footnote 158 though it has more recently ‘migrate[d]’ into international legal academiaFootnote 159 – that disasters are not acts of God, but a social construct involving the coalescence of natural hazards with human-induced vulnerabilities to create a ‘disaster’.Footnote 160 While ‘[t]he physical existence of disasters establishes an agency of nature that exists independently of human perception’, it is human society that ‘actualises the potential’ of natural hazards.Footnote 161 As Lauta writes:
[T]he scope of disasters has little to do with the hazard itself, and everything to do with the way we organise our societies.Footnote 162
From the sociological view that human activities and choices contribute to the impacts of a disaster, disasters have long been regarded as ‘amenable to regulation’,Footnote 163 both from a legal and moral standpoint.Footnote 164 Modern IDL is premised on the notion that positive action by states is required, and potentially effective, to prevent and mitigate disaster risks, and to enhance response and recovery efforts. This agenda has been advanced for several decades, most recently in instruments such as the Sendai Framework, which ‘internalises and operationalises’ the view of disasters as social constructs.Footnote 165 Indeed, this also explains the change in focus within the modern IDL programme from the ‘primarily response-centric model’ of earlier years to ‘one focused largely on prevention and preparedness’.Footnote 166
This evolution in thought and regulatory orientation has, for one, eroded the conceptual rationale of the technique of exculpation in general, i.e., human non-responsibility. This rationale is incongruent with the ‘mesh of liabilities and diffuse lines of responsibility’ that disasters create.Footnote 167 As Surat and Lezaun have remarked, ‘[t]he law is to function as an instrument of forensic scrutiny and produce clear lines of accountability where there was only confusion’.Footnote 168
More visibly, however, this also means that legal doctrines which reflect the anachronistic idea of disasters as rare and beyond human control may no longer be appropriate for disaster regulation. In particular, the availability of force majeure and necessity is conditional on a state not contributing to the situation.Footnote 169 Given modern sociological thinking, any state would be hard-pressed to argue that its policies on urbanization, emergency response and external assistance, among other things, had not contributed to the impacts of a disaster.Footnote 170 Though inadvertent contributions which were done ‘in good faith’ and which did not themselves make ‘the event any less unforeseen’ do not preclude reliance on force majeure,Footnote 171 it is difficult to sustain a characterization of disasters as unforeseen, nor cast systemic failures to enact disaster prevention measures as good faith omissions. Of course, this brings temporal issues into play, as it is not certain how far back one can go when assessing contribution before a policy becomes too remote.
Similarly, given that it is ‘extremely likely’ that human activity has been the ‘dominant cause’ of accelerated global warming since the mid-twentieth century,Footnote 172 which in turn has increased the likelihood and intensity of disasters, this may preclude reliance on force majeure on the basis that a state, by failing to adopt sufficiently ambitious emissions reduction policies, has assumed the risk of the situation occurring.Footnote 173
This conclusion equally applies to the requirement in VCLT Article 62 that a state seeking to suspend a treaty obligation on the basis of a ‘fundamental change in circumstances’ must demonstrate that the change was ‘not foreseen by the parties’ at the time of concluding the treaty. While parties brokering an investment treaty or free trade agreement, for instance, may not foresee specific disasters, Article 62 is phrased such that it is the change in circumstances which must not be foreseen, rather than the disaster itself. Therefore, it may be difficult to argue that they do not foresee that a fundamental change in circumstances, which radically transforms a key obligation to protect foreign investments or continue a free trade arrangement, might occur at any point by reason of a disaster, given their increasing statistical likelihood.
Conversely, however, doctrines which do embrace modern notions of human contribution to disasters remain better adapted. For instance, a broader conception of what is ‘necessary’ on the part of states to prevent disasters may render certain disapplication and exculpation mechanisms, e.g., NPM clauses and Article XX of the GATT, more easily invocable (their respective proportionality requirements notwithstanding). The heightened relevance of such doctrines can be explained on a conceptual level: such clauses, insofar they direct themselves not at disasters but at state responses to disasters, better encapsulate the notion of disasters as social constructs which humans must manage. In this way, they are also more consistent with the policy drivers and adaptive orientation of IDL. Therefore, state responses necessary to manage those risks (and in so doing, to protect ‘animal and plant life and health’ or ‘essential security interests’) fall within the clauses. The rationale for disapplication and exculpation in such clauses is not human non-responsibility for disasters, as it is for customary defences. Rather, it is the affirmative responsibility of states to manage disasters which justifies other conflicting obligations being disapplied, or the consequences of breach avoided.
Other mechanisms may also become more relevant because they better align with sociological and scientific realities. For one, the requirement for engaging VCLT Article 61 that there must be ‘permanent destruction’ of an object indispensable for the execution of the treaty may be more readily satisfied as climate change increases the intensity of disasters and their impacts on the many physical objects with which treaties are concerned, such as watercourses, forests, and nuclear facilities. For a similar reason, more disasters in future may satisfy the requirement for derogation under IHRL that a public emergency must ‘threaten the life of the nation’. Finally, international courts and tribunals may in future take a flexible interpretation of primary norms from other branches of international law when dealing with claims concerned with disasters, when IDL norms crystallize of sufficient particularity and bindingness to justify this approach.
In sum, certain doctrines embodying the regulatory techniques of disapplication and exculpation are ill-suited to regulate disasters, insofar as they depend on an understanding of disasters which has long been debunked. This unsuitability, first visible at a doctrinal level, reflects conceptual limitations in the reactive model for disaster regulation as a whole. This, in turn, explains the rise of adaptive regulatory initiatives which regard disasters as amenable to and requiring specific regulation. Nonetheless, there are some notable exceptions of reactive doctrines which remain well-adapted to modern disasters, including NPM clauses in trade and investment agreements, and derogation under IHRL. These doctrines retain their relevance because the disapplication or exculpation that they permit is grounded not on the belief that the disaster is beyond that state’s control, but that it is precisely within its control to prevent and manage.
5. Conclusion
This article has made two contributions which might inform future forays into the burgeoning field of IDL. First, in terms of its methodology, it has presented a theoretical framework by which to understand how disasters are reconciled with a state’s existing international legal obligations. This reactive model has been overlooked in IDL literature, at least in terms of its systemic relationship to the adaptive model that characterizes IDL’s modern complexion. It is hoped that the present inquiry serves to historicize and account for the emergence of this adaptive model; and offers a springboard to prompt further research into the orientation, structure, and epistemology of IDL as an academic discipline.
Secondly, in terms of the substantive material presented, it has argued that the techniques of disapplication and exculpation which are employed to reconcile disasters with existing obligations under international law are inadequate insofar as their legitimizing sociological and scientific premises do not pass muster. This is crucial to policymakers for two reasons. First, it validates the adaptive regulatory initiatives pursued by central IDL actors such as the ILC and the IFRC for their consonance with modern sociology and science. Secondly, it indicates a need to continue to develop and engage with disapplication and exculpation mechanisms that do align conceptually with modern understandings of disasters, and to recalibrate those that do not. Further research is warranted in this regard, both to identify precisely which aspects of each doctrine are and are not fit for purpose, having regard to real and hypothetical disasters, and to propose reforms on that basis. Efforts to address, balance, and integrate the reactive and adaptive models for disaster regulation are crucial for IDL to sustain claims to policy relevance.
