Foreign Surveillance and Human Rights, Part 3: Models of Extraterritorial Application

This post is part of a series: Intro, Part 1, Part 2, Part 4, Part 5.

In this post I will provide a brief outline of the (often conflicting and confusing) case law on the meaning of the concept of state jurisdiction in human rights treaties. I will examine the spatial model of jurisdiction, which conceptualizes it as effective overall control of an area, the personal model of jurisdiction as authority and control over individuals, and a third model which distinguishes between the positive and negative obligations of states under human rights treaties. My next post will proceed to apply these models to several possible factual scenarios of overseas surveillance.

The European Court has produced by far the most case law on extraterritorial application, both in quantity and in variety. No case that I am aware of, however, deals directly with the question of extraterritorial application of the Convention to foreign searches, interceptions, or surveillance. The issue is thus one very much of first impression. The jurisprudence of the Human Rights Committee, on the other hand, is not as conflicting or contradictory, even if it is less varied. The Committee has also generally been more generous towards applicants than the European Court; unless I am mistaken, there is no case in which the Committee rejected the communication of a person who made an arguable case that his or her rights were violated extraterritorially on the grounds that this person was not subject to the jurisdiction of the relevant state. (The Committee’s generosity can be explained, in my view, by the fact that it does not necessarily need to live with the consequences of an expansive approach in the same way as the Strasbourg Court, where the stakes are higher because of the greater robustness of the regime and the binding nature of the Court’s decisions).

The Spatial Model

The spatial model of jurisdiction as de facto effective control over areas is the least controversial. The European Court famously articulated it in the Loizidou case dealing with Northern Cyprus, the Human Rights Committee similarly applied it to e.g. the occupation by Israel of the Palestinian territories, and the ICJ likewise found the ICCPR to apply during occupation in the Wall and Congo v. Uganda cases. This approach makes intuitive sense: if a state exercises control over the territory of another state that in many replicates the extent of control it has over its own territory, then it is only appropriate for it to have human rights obligations towards the territory’s inhabitants. As the Court held in Loizidou, what matters is the fact of such control, regardless of whether it was obtained lawfully or unlawfully (i.e. in violation of the territorial state’s sovereignty).

The benefit of this conception of jurisdiction is its clarity. There will always be difficult or borderline cases (cf. Ilascu), but the test itself is still workable and provides some limits on states’ obligations. But the benefit is also its drawback, since the spatial model may be too limiting. There are many situations in which a state is factually perfectly capable of violating the rights of individuals without controlling the actual area – drones or ‘enhanced interrogation’ at CIA black sites being cases in point. And the more such cases keep popping up, the more morally arbitrary it seems to condition the state’s obligations on territorial control when such control is perfectly irrelevant for the substance of the violation.

One way of dealing with this problem is to shrink the size of the area that is the object of the effective overall control test. Northern Cyprus is surely such an area, but so is Guantanamo even though it is much smaller. There have also been a number of cases applying the spatial model to even smaller areas or places such as a British military prison in Iraq (Al-Saadoon), or man-made objects such as ships and aircraft.

Yet the more one shrinks the size of the area, the more artificial and arbitrary the whole test seems. For instance, should the application of human rights treaties really depend on whether state agents control the house in which an individual was shot to death, a possibility mooted during the litigation before the English courts in Al-Skeini? The more the size of the area shrinks, the more likely it is that the spatial model collapses into a conception of jurisdiction as control over individuals, rather than spaces.

The Personal Model

The idea that the word ‘jurisdiction’ in human rights treaties denotes authority and control exercised by states over individuals also has a long pedigree. It was first set out by the European Commission in one of the early interstate cases between Cyprus and Turkey, but its biggest proponent has been the Human Rights Committee. In Lopez-Burgos, a case dealing with an abduction by Uruguayan agents of an individual on Argentine territory, the Committee held that:

The reference in article 1 of the Optional Protocol to ‘individuals subject to its jurisdiction” does not affect the above conclusion because the reference in that article is not to the place where the violation occurred, but rather to the relationship between the individual and the State in relation to a violation of any of the rights set forth in the Covenant, wherever they occurred. … Article 2 (1) of the Covenant places an obligation upon a State party to respect and to ensure rights “to all individuals within its territory and subject to its jurisdiction’, but does not imply that the State party concerned cannot be held accountable for violations of rights under the Covenant which its agents commit upon the territory of another State, whether with the acquiescence of the Government of that State or in opposition to it. … In line with this, it would be unconscionable to so interpret the responsibility under article 2 of the Covenant as to permit a State party to perpetrate violations of the Covenant on the territory of another State, which violations it could not perpetrate on its own territory.

(emphasis mine)

Note how the Committee is essentially making an appeal to the universality of human rights in order to justify the personal model. It has reiterated this approach in General Comment No. 31, para. 10, when it held that ‘a State party must respect and ensure the rights laid down in the Covenant to anyone within the power or effective control of that State Party, even if not situated within the territory of the State Party … regardless of the circumstances in which such power or effective control was obtained.’

So articulated the principle is broad enough to make any human rights lawyer happy. But the benefit is again also a drawback, since it seems impossible to limit this principle in a non-arbitrary way. In other words, applying it consequentially would lead to the Covenant applying to any extraterritorial state action. While that may yet not overly concern the Committee, it was precisely this kind of fear – of possible overreach, lack of institutional competence, and all sorts of practical and political difficulties on the merits – that in the immediate wake of 9/11 led the European Court to render its Bankovic inadmissibility decision, which studiously ignored the personal model.

As the criticism of Bankovic mounted, as its arbitrariness got more and more exposed in a succession of smaller cases, and as the composition of the Court itself changed, the Court decided to systematically revisit the question of extraterritorial application in the Al-Skeini case (notably of the 17 judges who sat on the Al-Skeini Grand Chamber only 3 sat on the Bankovic Grand Chamber – Costa, Rozakis, and Casadevall, whereas Judge Costa, presiding over the Al-Skeini Grand Chamber, had also presided over the 2004 Issa Chamber which openly contradicted Bankovic and paraphrased the ‘unconscionable’ language of Lopez-Burgos).

In Al-Skeini the Court reaffirmed the validity both of the spatial model (‘effective control over an area,’ paras. 138-140) and the personal model (‘state agent authority,’ paras. 133-137), concluding with regard to the latter that (para. 137):

 It is clear that, whenever the State through its agents exercises control and authority over an individual, and thus jurisdiction, the State is under an obligation under Article 1 to secure to that individual the rights and freedoms under Section 1 of the Convention that are relevant to the situation of that individual. In this sense, therefore, the Convention rights can be “divided and tailored” (compare Banković, cited above, § 75).

The Court thus not only defined ‘jurisdiction’ as state ‘control and authority over an individual,’ but it also partly overruled Bankovic by allowing for the ‘dividing and tailoring’ of Convention rights, as opposed to the all-or-nothing Bankovic approach.  But the Court was still aware that if it defined jurisdiction in such terms the Convention would apply everywhere, as again there is no normatively sound, non-arbitrary way of concluding, for instance, that physical custody qualifies as ‘control and authority’, while killing (or the power to kill) does not. Indeed, in Al-Skeini the Court found that the five applicants killed by British troops on patrol in Basra were within the UK’s jurisdiction precisely because killing was authority and control (para. 149). The Court however found a limiting principle in the concept of ‘public powers’ that it imported from the Bankovic analysis of the spatial model of jurisdiction – the killing of the applicants was thus an exercise of UK jurisdiction, but only because due to the occupation of Iraq and relevant resolutions of the Security Council the UK ‘assumed authority and responsibility for the maintenance of security in South East Iraq.’

Al-Skeini was thus a major attempt by the Court at fixing Bankovic, in which it was partly successful. But it in fact preserved the result of Bankovic and by using the nebulous concept of ‘public powers’ managed to avoid the application of the ECHR to foreign military interventions simpliciter, as e.g. recently in Libya. (For more on this, see here). The uncertainties of Al-Skeini similarly left the door open for the UK to argue that it is confined to the specific facts of Iraq, and that the Convention largely does not apply to UK activities in Afghanistan.

A Third Model: Positive and Negative Obligations

Just like the spatial model can in extremis collapse into the personal one, the more the area subject to jurisdiction shrinks in size, so does the personal model ultimately collapse, with the extraterritorial application of human rights treaties becoming limitless. The European Court’s attempt to prop it up through the ‘public powers’ concept will work for a while, but will ultimately prove to be arbitrary and unstable.

I have hence argued in favour of a third model which would be based on the distinction between the overarching positive obligation of states to secure or ensure human rights, which extends even to preventing human rights violations by third parties, and the negative obligation of states to respect human rights, which only requires states to refrain from interfering with the rights of individuals without sufficient justification. Under this model, ‘jurisdiction’ would primarily mean effective overall control over areas, and the overarching positive obligation would be predicated on a state having such control over an area, because in the overwhelming majority of situations the state actually needs such control in order to be able to comply with this obligation.

On the other hand, the negative obligation to respect human rights would be territorially unlimited and not subject to any jurisdiction threshold. Textually this would flow from Article 1 ECHR only referring to the obligation to secure, while Article 2(1) ICCPR could reasonably be read as attaching the jurisdiction threshold only to the obligation to ensure, but not the obligation to respect. Alternatively, negative obligations could still be subject to the jurisdictional threshold under the personal model, but as we have seen this threshold actually collapses and the end result would be the same. The rationale for not limiting negative obligations is that states are always perfectly able of complying with them, since they remain in full control of their own organs and agents.

The moral logic of universality is thus brought to its conclusion, while jurisdiction still serves  as a limiting factor for the normally far more onerous positive obligations. I am not arguing that this model is perfect, but I do claim that it is clear, predictable, precludes the vast majority of arbitrary outcomes and provides a stable balance between considerations of universality and effectiveness. Similarly, while I argue that this is how human rights treaties should be interpreted, I am not claiming that this is what human rights bodies or courts already are doing. Rather, I am saying that this model presents a stable equilibrium towards which the spatial and personal conceptions of jurisdiction will naturally tend to gravitate.

My next post will apply these models of jurisdiction to several possible scenarios of overseas surveillance.