COVID-19 and Defences in the Law of State Responsibility: Part II

Written by and

In our previous post, we considered whether States could rely on the plea of force majeure in respect of non-performance of international obligations connected to their efforts to contain the COVID-19. We concluded that force majeure might not provide a defence to States since their measures in addressing the spread of the virus were voluntary measures. In this post, we consider whether States could rely on the pleas of necessity or distress in the same circumstances.

State of necessity

Article 25 of the ARS states that:

1. Necessity may not be invoked by a State as a ground for precluding the wrongfulness of an act not in conformity with an international obligation of that State unless the act:

(a) is the only way for the State to safeguard an 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.

2. In any case, necessity may not be invoked by a State as a ground for precluding wrongfulness if:

(a) the international obligation in question excludes the possibility of invoking necessity; or

(b) the State has contributed to the situation of necessity.

To successfully plead the defence of necessity, a State must fulfil the following four requirements: (i) there must be a grave and imminent peril; (ii) this peril must threaten an essential interest; (iii) the State’s act must not seriously impair another essential interest; (iv) the State’s act was the ‘only way’ to safeguard the interest from that peril. In addition, the plea is excluded if: (v) the obligation in question excludes reliance on necessity; and (vi) the State contributed to the situation of necessity. These will be addressed in turn, with the exception of (v) as this will depend on an interpretation of the specific obligation impaired.

First, there must be a grave and imminent peril. This requirement is best understood by separating the elements of risk and harm. What is required is that there is an imminent risk that an essential interest will be gravely harmed. The peril (the risk) must be ‘imminent’ in the sense that it has not materialised (Commentary, §16). Further, imminence does not mean immediacy (as  in the peril is about to happen), but rather refers to certainty: can it be established, with a (sufficient) degree of certainty that the peril will occur at some point in the future? Likewise, the harm need not have occurred at the time a State acts in a situation of necessity – after all, the point of the defence is to prevent that harm from occurring. But the defence should also be available where States aim to prevent an unfolding harm from becoming graver or more extensive (though it may be queried whether, by not acting before the harm begun to materialise, the State has contributed to the situation of necessity).

In short, the risk must be ascertained now that an essential interest will suffer harm (or continue to suffer harm) in the future. The source of the harm (that is, the event that will cause harm to the essential interest) can be a past event, an ongoing event, or an event in the future. And the risk can refer to the occurrence of the event itself (so the State can act to prevent its occurrence) or to the materialisation of the harm as a result of the event (so the State can adopt measures to safeguard interests from the event). A ‘measure of uncertainty about the future’ does not necessarily exclude the plea, so long as the peril can be ‘clearly established on the basis of the evidence reasonably available at the time’ (Commentary, §16).

The outbreak and spread of COVID-19 would appear to meet this requirement. It is an unfolding event which poses an imminent threat of a grave harm to the world’s population. The basic reproduction number (R0) is agreed to be over 1, meaning that on average every infected individual is likely to transmit the virus to more than one other individual, on the assumption that, as appears to be the case here, the population is susceptible to infection (it is unlikely that anyone was immunologically protected from the virus prior to the epidemic starting, so it is safe to assume that everyone is susceptible to infection). The current case fatality ratio (CFR) estimates, namely the estimated proportion of people infected who are likely to die from the disease, are almost unanimously above 0.5%. There are estimates that for particular sections of the population the CFR is far higher (see here, here and here). These estimates, however, remain tentative due to, among others, the fact that asymptomatic cases or patients with mild symptoms might not be tested and, therefore, will not be identified. The fast spread of the infection, coupled with the potentially high and as yet uncertain mortality rate, poses a grave harm to the population.

Second, the peril must threaten ‘an essential interest’. Article 25 does not require that this be an essential interest of the State invoking the plea, so it could be an essential interest of the international community or, potentially, of other States. The Commentary explains (§15) that whether an interest is essential is a relative question and that the interest relied on ‘must outweigh all other considerations … on a reasonable assessment of the competing interests, whether these are individual or collective’ (§17). As evidenced by the decision of the WHO Director-General to classify the outbreak as a Public Health Emergency of International Concern, the outbreak of COVID-19 poses a serious threat to the health and lives of individuals within a State but, crucially, also in other States. It also poses a threat to the continued functioning of health services. By way of example, the well-being of a States’ population and the continued functioning of its public services have been accepted as constituting ‘essential interests’ in investment treaty arbitration (see National Grid v Argentina, §245). Of course, this determination is affected by the character of the interest impaired, but it seems safe to assume that in most circumstances, the well-being of individuals and, in particular, the safeguarding of their lives will be an essential interest.

Third, the act must not ‘seriously impair’ an essential interest of another State or of the international community as a whole. As already noted, investment tribunals have readily accepted that a State’s interest in the wellbeing of its population outweighs the interests of investor home States (or those of investors themselves). The balance may be differently struck when the interests impaired are protected by human rights law (should the plea of necessity be applicable). Quarantines and isolation measures restrict, among others, individuals’ freedom of movement and of assembly (on which see here). It is arguable that the measures do not seriously impair the interests protected by these rights, since they are only temporary measures and these are rights which are subject to derogation in most human rights systems. This said, each case will need to be assessed independently, by reference to the relevant interests involved.

Fourth, the act of the State must be the ‘only way’ to protect the essential interest from the impending harm at the time the State adopts the relevant measures. If there are other (lawful) ways to address the threat, even if these are more costly or inconvenient, the plea will fail. The assessment is to be made from the standpoint of the State organs adopting the measures – with the degree of knowledge available to them at that time. It is crucial not to allow hindsight biases, which are likely to arise once we know what measures were effective, to colour the assessment. Protecting the essential interest in question (namely, the lives of individuals, and their wellbeing through the continued functioning of public healthcare services), requires either containing the disease (reduce to 0 the number of infections) or mitigating its effects (e.g. controlling the spread of the disease, to ensure adequate access to healthcare provision for those who will develop severe symptoms). Given all the uncertainties in respect of the virus at the time it broke out (many of which remain relevant today), and the lack of effective vaccines or targeted treatments available to States, some form of extreme social distancing measures (quarantining, banning of mass gatherings, etc) was all that States could do to mitigate the morbidity and mortality of the virus within their populations.

Epidemiologically, there is no single way to address these viral epidemics. The two main approaches, containment and mitigation, will require social distancing measures – either for individuals who have been exposed to the virus, or for the population at large. Social distancing measures can be very limited, like requiring self-isolation of those potentially exposed to the virus, to more extreme, like the lockdowns imposed in places like Italy. But the implementation of these measures can (and has) varied in different States. These variations depend on  a number of factors, including whether the virus is believed to have already reached a State’s territory, the stage of the epidemic in a State’s own territory,  the situation in other States (for example, if there are no cases in other States then there is no longer a risk of introduced cases and measures such as airport screening would become redundant), and what may be culturally and logistically viable within that State’s population (for example, some States may not have adequate resources to consider enforcing large quarantines). In the absence of an effective vaccine or targeted treatment, a package of these measures (taken together) are, on the whole, the ‘only way’ to contain or mitigate the spread of the disease. However, when looked at individually, some or other measures may turn out not to be the ‘only way’. Travel bans, for example, are unlikely to be the only way (or a reasonable way, even) to deal with the epidemic once the virus has become widespread within a territory (see here and here). There is also good mathematical modelling showing that entry screening of people coming from affected countries into unaffected countries only delays the start of the epidemic in the unaffected country, but that this effect diminishes quickly as the number of infected individuals entering increases (see here and here). Whether a State is able to meet the ‘only way’ requirement will ultimately turn on the obligation allegedly breached and whether the assessment is made by reference to the package of measures, as a whole, or to each individual measure.

Finally, the plea may not be relied upon if the State has ‘contributed’ to the situation of necessity. The State’s contribution must be ‘sufficiently substantial and not merely incidental or peripheral’ (Commentary, §20). This standard is more ‘categorical’ than the equivalent standard in Articles 23 (force majeure) and 24 (distress). As in the case of force majeure, there is considerable uncertainty as to what this standard requires. Some tribunals have approached the requirement as a purely causal one such that ‘well-intended but ill-conceived policies’ are sufficient to exclude reliance on the plea (Impregilo v Argentina, §356). Others have interpreted it more narrowly, as requiring some degree of fault (most recently, Urbaser v Argentina, §711). Furthermore, there may be temporal issues as well: at what point in time do policies and other State behavior become too remote to be a relevant contribution to the situation? On a purely causal understanding, it would be arguable that States’ limited funding of public healthcare services, say as a result of policies of austerity adopted after 2008, had contributed to the crisis: quarantines and social distancing measures are, among others, intended to delay and spread, over time, the rate of infections so as to ensure adequate access to existing healthcare provision. If there were adequate healthcare provision, such measures may not have been needed. In contrast, if the standard of contribution were narrower, perhaps requiring negligence or some other degree of fault, States might be able to argue that good faith policies, even if they affected funding in public healthcare, could not exclude reliance on the plea.

The above analysis suggests that the argument based on state of necessity is more promising than that based on force majeure, although there may still be some difficulties. Indeed, the plea has been interpreted in very restrictive terms by tribunals and States may struggle to prove that their measures were the ‘only way’ and that their prior behavior did not contribute to the situation of necessity.


An argument based on distress could be the most successful. Article 24 ARS states that:

1. The wrongfulness of an act of a State not in conformity with an international obligation of that State is precluded if the author of the act in question has no other reasonable way, in a situation of distress, of saving the author’s life or the lives of other persons entrusted to the author’s care.

2. Paragraph 1 does not apply if:

(a) the situation of distress is due, either alone or in combination with other factors, to the conduct of the State invoking it; or

(b) the act in question is likely to create a comparable or greater peril.

To make this plea, the State would need to prove that: (i) threat to life; (ii) a special relationship between the State organ and the persons in question; (iii) that there was no other reasonable way to deal with the threat; (iv) that it did not contribute to the situation; and (v) that the measures were proportionate.

As to the first requirement, this should be easily met. As noted earlier, there is a threat to the lives of individuals, especially vulnerable sections of society.

Second, there must be a ‘special relationship’ between the individuals in question and the State organ adopting the measure. This requirement is intended to exclude the application of the plea to situations of ‘more general emergencies’ which, according to the Commentary, are ‘more a matter of necessity than distress’ (§7). There may be some difficulties fulfilling this particular requirement. A general relationship between the State and its own population seems to be insufficient. The examples given in the Commentary suggest that one important aspect of the ‘special relationship’ between the organ making the decision and the individuals whose lives are under threat is control: is the fate of those individuals under the control of the relevant organ? This aspect can readily be perceived in situations involving ships or aircraft in distress: only a decision by a ship-master can avert the threat to the passengers’ lives. But it does not seem necessary that this control be physically close. In Rainbow Warrior the decision to repatriate French agents Mafart (§17) and Prieur (§55) was taken by the central French government and not, for example, by the local officers in charge of their wellbeing. Insofar as only the central government has the authority to put in place measures of containment or mitigation in these types of emergencies, it is plausible to argue that there is a special relationship: to some extent, the fate of the population is within the control of the central authorities. An argument is therefore plausible that the special relationship criterion is met.

Third, the plea of distress requires that the measures are the only ‘reasonable way’. The reasonableness criterion is intended to provide some ‘flexibility regarding the choice of action’ (Commentary, §6). This is a lower standard than for the plea of necessity which required that the measures be the ‘only way’ to deal with the emergency. Social distancing measures are a reasonable avenue, in view of the absence of vaccines or targeted treatments. However, it is important once again to emphasise that the package of measures, and perhaps each measure, adopted by each State may need scrutiny. Taken individually, some measures, like travel bans once there is within-country transmission, may not be reasonable.

Fourth, the non-contribution requirement is here narrower than in the case of necessity: good faith policies that contributed to the crisis do not exclude reliance on the plea. The standard is the same as in force majeure, but the Commentary adds that ‘priority should be given to necessary life-saving measures’ (§10). This seems reasonable, as it would be illogical to prevent the saving of lives because the State had contributed to the situation. In this analysis, as we noted in the contexts of force majeure and state of necessity, much will turn on how the standard is interpreted: whether fault, and in what degree, is necessary; whether temporal proximity is required, etc. This said, it is possible that decision-makers will be less stringent in the application of this standard if they interpret it teleologically, by reference to the need to save lives.

Last, the measures must not create a comparable or greater peril. This is a proportionality requirement between the measures adopted, on the one hand, and the interest protected by the obligations impaired. The interest protected, like in the case of necessity, must ‘clearly outweigh the other interests at stake in the circumstances’. This element is likely to be met, in line with our analysis in the context of the plea of necessity.


None of the three defences in the law of State responsibility are applicable with certainty to every State, let alone to all States. Much will depend on the specific obligations allegedly breached, the specific measures impugned, and the timing of the epidemic. Force majeure is likely to be the least relevant, since the State behaviour which is potentially in breach of international obligations is voluntary: these are measures that States have decided to put in place to contain or mitigate the crises. While this is unthinkable, States could choose not to do anything about it and continue to operate business as usual. State of necessity and distress are potentially more suitable, but in both cases there are elements which may be difficult to satisfy. In the case of necessity, the ‘only way’ requirement is likely to fail as there is no scientifically agreed measure to deal with viral epidemics such as this. For distress, there will be some difficulty (but presumably less so) in satisfying the ‘special relationship’ requirement. In both cases, moreover, the non-contribution requirement, given its vagueness, can be an obstacle.

These defences were drafted by the ILC, with the support of States in the Sixth Committee of the UNGA, in very stringent terms, precisely to limit as much as possible States’ ability to rely on them given that, in most cases, States will make these decisions unilaterally and in contexts in which there is no settlement body with mandatory jurisdiction over disputes potentially arising from these situations. A balance was struck between the importance of ensuring State compliance with their obligations, on the one hand, and the need for flexibility in addressing unexpected circumstances, on the other. Yet, none of them seem to adequately cater to the present situation – a global emergency, as attested by the WHO, if there ever was one. It is perhaps not unsurprising that these defences are somewhat out of step with current threats: they were drafted on the basis of past practice which was, sometimes, centuries old. The scientific community no longer conceives of these events as natural disasters: these are man-made emergencies, which are the result of a combination of causes, including natural events, long term ecological and political trends, as well as human and societal choices.

The defences in the law of responsibility, as evidenced by this example, appear to be unable to cater to these understandings. The non-contribution requirements, in particular, tie the hands of States: if emergencies are understood to result from, among others, their own policies, then rarely will they be able to rely on defences such as necessity (specifically drafted to address emergencies), unless these requirements are very narrowly circumscribed. A new balance between the needs for compliance and flexibility built into these rules may need to be struck so as better to reflect the scientific understanding of these emergencies. States should not unjustifiably breach obligations owed to other international actors, but at the same time they should have sufficient flexibility to address these disasters – especially when the lives of individuals, both within their jurisdictions and beyond, are at stake.

Print Friendly, PDF & Email

Leave a Comment

Comments for this post are closed