Georgia v. Russia No. 2: The European Court’s Resurrection of Bankovic in the Contexts of Chaos

Written by

Last week the Grand Chamber of the European Court of Human Rights delivered its judgment in the second interstate case brought by Georgia against Russia (no. 38263/08), dealing with the August 2008 conflict between the two states (see my brief preview here ; for a summary of the judgment see the Court’s press release here). Briefly, the Court found Russia responsible for serious human rights abuses in the immediate aftermath of the conflict. But, by 11 votes to 6, it refused to look at any of the alleged substantive violations of the right to life during the ‘active hostilities’ phase of the conflict.

This is a very important judgment – a substantial, but far from total win for Georgia – but also a very disappointing one. Regardless of the final outcome, the Court’s reasoning on the key points of jurisdiction and the applicability of the European Convention in armed conflict is exemplary only in its arbitrariness. It is a retrograde step, putting the Court firmly against the overarching trend in international jurisprudence and underscoring its position as an outlier when compared to other human rights bodies. For example, the Court does not even cite, let alone seriously engage with, the Human Rights Committee’s General Comment No. 36 on the right to life, or the relevant jurisprudence of the Inter-American Commission and Court.

The one good thing that can be said about the judgment is that the Court’s policy-driven motivations behind it are entirely transparent. Unlike many of us in the academia, who have seen the current unprecedented number of interstate cases as an indicator of a ‘golden age’, the Court – or at least this split constellation of the Court, fearful of political backlash – sees such cases dealing with armed conflict as an existential danger, which it wants to avoid dealing with as much as it is possible to do so. Thus the approach that the Court took here, a retreat from Al-Skeini back to Bankovic (and more), does not bode well for many of the similar pending cases, such as those on Ukraine or the Armenia/Azerbaijan/Turkey litigation re Nagorno Karabakh.

The Court’s judgment in brief

But let’s take this slowly. This is a very long and complex judgment. It needs some unpacking. To understand the case, the key point is a temporal one – the distinction between the five-day ‘active hostilities’ phase of the conflict, 8-12 August 2008, and human rights violations that happened in that period, and the occupation phase after the EU-mediated ceasefire on 12 August, and subsequent violations. This distinction shaped the whole judgment.

With respect to events after 12 August, the Court found that Russia was in effective control of the separatist entities of Abkhazia and South Ossetia, and also of the ‘buffer zone’ established by Russian armed forces on the territory of Georgia proper. It applied a spatial notion of jurisdiction in the sense of Article 1 ECHR, as control over territory, and found that Russia had control over all of these areas, exercised directly by its own forces or by the proxy armed groups. By and large its reasoning in this regard is straightforward and unproblematic. It is of course fact- and evidence-dependent – a good parallel being the recent admissibility decision re Crimea. The Court was nearly-unanimous on this jurisdiction issue (16 to 1).

It is the Court’s approach to the ‘active’ phase of the conflict that is deeply troubling, as is indeed the whole idea that the ‘active’ phase and the occupation phase can be separated so rigidly. Here the Court, by 11 votes to 6, found that it could not apply either the spatial model of jurisdiction (control over territory) or the personal model (authority, control over the victim by a state agent) to any of the alleged substantive violations of Article 2 ECHR (right to life) committed in that period. This is where the Court’s reasoning, which we will explore in detail momentarily, reaches the apex of arbitrariness.

But the Court’s approach is in the final analysis less restrictive than it might initially appear. First, the restrictive approach is limited to violations of Article 2, i.e. to (not) assessing the lawfulness of kinetic uses of force during the conflict – to an extent the status quo under prior ECHR jurisprudence, primarily Bankovic. But when it comes to the rights of individuals who were detained and subjected to ill-treatment even during the active phase of the conflict, the Court did find that Russia had jurisdiction over these persons even before 12 August.

Second, the Court’s Bankovic-inspired restrictive jurisdictional approach to kinetic uses of force is limited only to the substantive aspect of Article 2, i.e. the duty to respect the right to life by refraining from unjustified uses of force. But what the Court took away with one hand it gave back with the other, by saying that Russia had the procedural obligation under Article 2 to investigate potentially unlawful uses of lethal force committed during the active phase of the conflict. Its approach to jurisdiction in that regard is hardly a model of clarity, as we will see, but again it mitigates somewhat the restrictiveness regarding the substantive obligation and has major implications for foreign military interventions by European states. And its holding on both the detention issue and the procedural duty to investigate was unanimous.

So, to sum up, the Court found that Russia:

  • Did not have Article 1 jurisdiction under either the spatial or the personal model for the purpose of assessing the lawfulness of kinetic uses of force, such as artillery shelling, during the active phase of the conflict.
  • Did have jurisdiction, however conceptualized, for the purpose of investigating potentially unlawful uses of force even during the active phase of the conflict.
  • Did have spatial jurisdiction after 12 August 2008, up to the moment of the withdrawal of its troops several months later.
  • Did have jurisdiction (but it’s unclear on what basis) over all persons who were deprived of their liberty, even during the active phase of the conflict.

In short, the protective vacuum that the Court opens up here is the same vacuum that has already existed in its case law post-Bankovic ­– kinetic uses of force without territorial control. What this case adds to Bankovic is that the uses of force in question were not only aerial bombardment (although there was that too), but include ground combat. But even there the addition does not count for much, because distinguishing between aerial bombing and long-range artillery shelling would make very little sense in any event. And, just as with Bankovic, what makes the judgment so unpalatable is not simply the basic bottom line that the Court reaches, but the reasoning it employs to do so. So let’s explore that reasoning in more detail.

(Health warning: this is a very long post).

Jurisdiction and kinetic use of force in international armed conflict

At the outset of its jurisdiction analysis, the Court confines its holding (and continues doing so throughout) to international (i.e. interstate) armed conflict (IAC):

  1. In the present case the Court is required to examine whether the conditions applied by the Court in its case-law to determine the exercise of extraterritorial jurisdiction by a State may be regarded as fulfilled in respect of military operations carried out during an international armed conflict.
  2. In that connection it can be considered from the outset that in the event of military operations – including, for example, armed attacks, bombing or shelling – carried out during an international armed conflict one cannot generally speak of “effective control” over an area. The very reality of armed confrontation and fighting between enemy military forces seeking to establish control over an area in a context of chaos means that there is no control over an area. This is also true in the present case, given that the majority of the fighting took place in areas which were previously under Georgian control. (emphasis added)

The IAC point aside, the question of whether continued, intense fighting precludes a finding of effective control over territory is a perfectly valid one. Indeed this was precisely the issue in Al-Skeini, where the House of Lords found that the intensity of the insurgency in southern Iraq meant that the UK was not really in control over the territory, despite nominally being the occupying power. That’s, again, a perfectly defensible position.

The problem with the Court’s reasoning here is its categorical nature. It’s one thing to say that control over territory is a very fluid thing, especially in a very short time-frame and with respect to a relatively small area of territory, as on the facts of this case. But it is in principle perfectly possible for an invading army to gradually establish reasonably stable control over areas of enemy territory, as it advances through it, even though the hostilities may still be happening on the fringes of that territory. Think only of the invasion of Iraq by the US and UK in 2003 – it’s not as if control over Iraqi territory by the invading forces was some kind of binary on/off, all or nothing switch. That said, if the Court’s reasoning here was given a more nuanced, less categorical interpretation in light of the specific facts of the case, the conclusion that Russia was not in control of the areas in which the fighting was taking place would be at least plausible.

(By the way, is it just me or wouldn’t ‘Contexts of Chaos’ be a great name for a heavy metal band? But I digress.)

The Court therefore finds that the spatial concept of jurisdiction cannot be applied, and that it should look at the case from the standpoint of the personal model (para. 127). It recapitulates a number of cases in that regard that were contradictory to Bankovic in that they found that the mere use of kinetic force was sufficient to create a jurisdictional link, such as Issa or Pad (128-30), and attempts to distinguish them in a wholly unconvincing way:

  1. Admittedly, in other cases concerning fire aimed by the armed forces/police of the States concerned, the Court has applied the concept of “State agent authority and control” over individuals to scenarios going beyond physical power and control exercised in the context of arrest or detention.
  2. However, those cases concerned isolated and specific acts involving an element of proximity.
  3. By contrast, the active phase of hostilities which the Court is required to examine in the present case in the context of an international armed conflict is very different, as it concerns bombing and artillery shelling by Russian armed forces seeking to put the Georgian army hors de combat and to establish control over areas forming part of Georgia. (emphasis added)

So, the Court is first forced to admit that the personal conception of jurisdiction cannot logically be confined only to situations of arrest and detention – see in that regard the judgment of Mr Justice Leggatt in Al-Saadoon, who sensibly points out that a ‘principled system of human rights law [cannot] draw a distinction between killing an individual after arresting him and simply shooting him without arresting him first.’ But then the Court says that these cases are distinguishable because they ‘concerned isolated and specific acts involving an element of proximity.’ Even if interpreted charitably, this is an extraordinarily vague statement. What can this even mean?

One possible reading is that the Court here appears to be saying that some kind of one-off use of lethal force – be it an extraterritorial assassination such as that of Litivinenko or the Salisbury attacks, or even potentially drone strikes – is somehow more deserving of protection than a massive, systematic use of lethal force. But how would that be justified? Another reading is the focus on proximity, i.e. that somehow the situation of a soldier shooting somebody with a gun or rifle, or stabbing them with a knife, is more deserving of protection than one of distant bombing or artillery shelling (or say a cyber operation against a hospital). But again, this seems wholly unconvincing as a principled boundary of jurisdiction. You can bet, however, that future litigation in this area will zero in on the implications of this particular paragraph.

Then the Court, invoking Bankovic, reaches its bottom line (137-8):

the Court attaches decisive weight to the fact that the very reality of armed confrontation and fighting between enemy military forces seeking to establish control over an area in a context of chaos not only means that there is no “effective control” over an area as indicated above, but also excludes any form of “State agent authority and control” over individual … the conditions it has applied in its case-law to determine whether there was an exercise of extraterritorial jurisdiction by a State have not been met in respect of the military operations that it is required to examine in the instant case during the active phase of hostilities in the context of an international armed conflict.

This second appearance of the Context of Chaos is obviously more impactful, and is followed by the categorical statement that any form of personal jurisdiction is excluded during the active phase of an IAC. The Court then supports this position (139) by pointing to the lack of extraterritorial derogations in IACs, but as explained elsewhere that is really neither here nor there for conceptualizing jurisdiction under Art. 1 of the Convention. Note, however, that the Court’s position is actually less categorical than it may first seem – not only because of the vague (and easily stretchable) references to proximity above, but also because later in the judgment it expressly finds that persons detained during the active phase of hostilities were covered by the Convention. This means that situations like special forces operations, close combat not involving artillery/aerial bombing, ‘proximate’ killing of persons hors de combat, e.g. the wounded on the battlefield, may well be covered by the personal notion of jurisdiction as understood by the Court here.

Finally, the Court concludes its analysis by saying what it really means with all this jurisdiction stuff:

  1. That said, the Court is sensitive to the fact that such an interpretation of the notion of “jurisdiction” in Article 1 of the Convention may seem unsatisfactory to the alleged victims of acts and omissions by a respondent State during the active phase of hostilities in the context of an international armed conflict outside its territory but in the territory of another Contracting State, as well as to the State in whose territory the active hostilities take place.
  2. However, having regard in particular to the large number of alleged victims and contested incidents, the magnitude of the evidence produced, the difficulty in establishing the relevant circumstances and the fact that such situations are predominantly regulated by legal norms other than those of the Convention (specifically, international humanitarian law or the law of armed conflict), the Court considers that it is not in a position to develop its case-law beyond the understanding of the notion of “jurisdiction” as established to date.
  3. If, as in the present case, the Court is to be entrusted with the task of assessing acts of war and active hostilities in the context of an international armed conflict outside the territory of a respondent State, it must be for the Contracting Parties to provide the necessary legal basis for such a task.
  4. The Court reiterates in this connection that this does not mean that States can act outside any legal framework; as indicated above, they are obliged to comply with the very detailed rules of international humanitarian law in such a context.

Translated: we’re oh so very sorry (‘may seem unsatisfactory to the alleged victims’), but war is not our problem. Let others try to solve it – states, IHL, the UN, international criminal courts – but not us. Even if we did try, it wouldn’t do us any good – certainly not in IACs – and us trying might harm the Convention system as a whole. (Note again how the Court repeatedly confines its holding to IACs only, leaving it entirely unclear how it would treat situations in which there is an (extraterritorial) NIAC, or lethal force is used in circumstances below the armed conflict threshold.)

I actually have a great deal of sympathy for this position. The sentiment is genuine and honest. The difficulties are very real. And I have long argued that cases such as these are not determined by the supposed intricacies of the notion of ‘jurisdiction’, but by a fundamental apology v. utopia tension between considerations of universality (‘doing the right thing’) and effectiveness (‘we’re not best placed to do this, and if we try the whole thing may blow up in our face’).  It is exactly this tension that drove Bankovic – as I tell my students, the one thing they need to understand about that case is the date on which it was decided (December 2001). So it is here. The Court – or at least this configuration of the Court – does not really want to deal with these cases, where it feels there is little good it can actually do, while the danger of political blow-back detrimental to the whole Convention system is too high to ignore. But, as I have also argued, the better way of dealing with these difficulties, including the relationship between the Convention and IHL, is to address these hard questions on the merits, rather than by distorting the notion of Art. 1 jurisdiction and turning it into a disguised, opaque and arbitrary non-justiciability doctrine. At least here, unlike in Bankovic, the Court was more open about what it was really doing.  

Spatial jurisdiction after 12 August 2008

Then, the Court shifts gears to the period after the ceasefire on 12 August 2008. And it has no great difficulty in establishing that Russia exercised control over Abkhazia, South Ossetia, and the ‘buffer zone’ in Georgia proper (161-75). The factual assessment is rather brief, focusing on the strength of the Russian military presence, the support it had given to the two separatist entities, relying mostly on various international reports (just like the recent Crimea decision). By and large this is entirely unobjectionable, except for the binary, on/off conceptualization of spatial jurisdiction. Can we really seriously accept that Russia was not in control of (some of) these areas on 11 August, but somehow was in (full) control of all of them on 12 August?

It is simply impossible to predict what the Court would do with this approach in case of situations that are even more fluid, e.g. Eastern Ukraine or Nagorno Karabakh and adjacent territories. And of course the fact that the fighting had ceased doesn’t mean that the IAC between the two states had ended – yet the Court sees no problem in looking at violations after the cut-off date.

The Court here makes one important further obiter remark, on an issue not previously explored in its case law, but as noted above directly raised by the House of Lords in Al-Skeini – the relationship between the threshold of belligerent occupation in IHL and the spatial notion of jurisdiction under Article 1 ECHR:

  1. Generally speaking, international humanitarian law applies in a situation of “occupation”. In the Court’s view, the concept of “occupation” for the purposes of international humanitarian law includes a requirement of “effective control”. If there is “occupation” for the purposes of international humanitarian law there will also be “effective control” within the meaning of the Court’s case-law, although the term “effective control” is broader and covers situations that do not necessarily amount to a situation of “occupation” for the purposes of international humanitarian law.

It’s debatable whether this approach is the right one – it is likely overly simplistic, even if it is correct in its essence. That every occupation equals Article 1 jurisdiction presupposes a clear line in IHL terms between an invasion and an occupation (there often isn’t any), and neglects the Al-Skeini issue of subsequent insurgency within an occupied territory. And then the whole idea of occupation necessarily requiring effective control has been very contested in recent years, most importantly with regard to Israel and Gaza (see also here). But again, the main conclusion that Russia was in control of the relevant areas is likely the correct one on the facts.

Attribution of the conduct of South Ossetian forces to Russia

However, another unfortunate feature of this judgment is that it is – yet again – unclear whether the Court is finding that the conduct of the South Ossetian organized armed groups, here violence against ethnic Georgians and the burning of their homes, was attributable to Russia. In other words, while Russia is being found responsible for an ‘administrative practice’ in that regard, it is unclear whether Russia is being held responsible for violating (by action) a negative duty to respect human rights, or for violating (by omission) a positive duty to prevent third parties from violating human rights within an area under its jurisdiction. This is a long-standing problem. The Court does not refer to the attribution tests developed by the ICJ and the ILC, and what it says could be interpreted either way (214, 217-8):

However, in accordance with the Court’s case-law, from the time when the Russian Federation exercised “effective control” over the territories of South Ossetia and the “buffer zone” after the active conduct of hostilities had ceased, it was also responsible for the actions of the South Ossetian forces in those territories, without it being necessary to provide proof of “detailed control” of each of those actions. …  In the present case, even if some witness statements indicate that at times Russian troops had intervened to stop abuses being committed against civilians, in many cases Russian troops were passively present during scenes of looting. … This shows that despite the order given to the Russian armed forces to protect the population and carry out peacekeeping and law-enforcement operations “in South Ossetia and in the Georgian territory controlled by the Russian armed forces” (see paragraph 189 above), on the ground the measures taken by the Russian authorities proved to be insufficient to prevent the alleged violations.

The same lack of clarity exists with regard to the treatment of persons in detention:

  1. Even if the direct participation of the Russian forces has not been clearly demonstrated, since it has been established that the Georgian civilians fell within the jurisdiction of the Russian Federation, the latter was also responsible for the actions of the South Ossetian authorities, without it being necessary to provide proof of “detailed control” in respect of each of their actions (see paragraph 214 above).

And with regard to the treatment of prisoners of war:

  1. Even if the direct participation of the Russian forces has not been clearly demonstrated in all cases, since it has been established that the prisoners of war fell within the jurisdiction of the Russian Federation on account of the “strict control” that it exercised over the South Ossetian forces, it was also responsible for the latter’s actions, without it being necessary to provide proof of “detailed control” of each of those actions (see paragraph 214 above).
  2. Furthermore, it can be seen from the above-mentioned reports and the statements of the Georgian witnesses that Russian forces were present on site and that they did not intervene to prevent the impugned treatment. (emphasis added)

Look carefully at para. 276 and the reference to ‘strict control’ in quotation marks, which does not appear in paras. 214, 248 or indeed (despite the quotation marks) anywhere else in the judgment. Note also how para. 276 refers to Russian control over South Ossetian forces, which looks like an attribution inquiry rather than as a jurisdictional one (control over an area). I may be making a wild guess here, but my hunch is that this stray ‘strict control’ language here is likely a leftover from an earlier draft of the judgment which did try to distinguish more clearly between the attribution and jurisdiction issues, and between Russia’s responsibility for violating negative and positive obligations. But this is just a guess, I may well be completely wrong.

Jurisdiction over persons in detention

As noted above, the Court found that all of the individuals detained by Russian or pro-Russian forces were within Russia’s jurisdiction, even if they were detained before 12 August 2008. But it’s worth reflecting on just how laconically – even furtively – the Court reaches that conclusion. Thus, with respect to Georgian civilians who have been detained, it finds that:

  1. In so far as the Georgian civilians were mostly detained after the hostilities had ceased, the Court concludes that they fell within the jurisdiction of the Russian Federation within the meaning of Article 1 of the Convention (see paragraph 175 above) and dismisses the preliminary objection raised by the respondent Government in that regard. (emphasis added)

That one word, ‘mostly,’ does most of the work here, covering all of the detained civilians. And the Court does the same thing for the detained prisoners of war:

  1. Given that they were detained, inter alia, after the cessation of hostilities, the Court concludes that they fell within the jurisdiction of the Russian Federation for the purposes of Article 1 of the Convention (see paragraph 175 above) and dismisses the preliminary objection raised by the respondent Government in that regard.

Again, note the cheeky ‘inter alia’ doing the work here. The fact that all detained civilians and PoWs undoubtedly were within the respondent’s jurisdiction becomes apparent upon reading operative paragraphs 4 and 8 of the judgment, which clearly refer to the period of active fighting before 12 August (‘civilians detained by the South Ossetian forces in the basement of the “Ministry of Internal Affairs of South Ossetia” in Tskhinvali between approximately 10 and 27 August 2008 fell within the jurisdiction of the Russian Federation’ and ‘Georgian prisoners of war who were detained in Tskhinvali between 8 and 17 August 2008 by the South Ossetian forces fell within the jurisdiction of the Russian Federation’)

This is of course the right result. It would be completely arbitrary to say that in a detention facility brimming with people only those who were detained after 12 August were protected by the Convention, but those detained earlier were not. Which is why the Court ‘mostly’, ‘inter alia’ doesn’t do that. But what the Court also doesn’t do is simply refer to its previously crystal-clear position in Hassan that any person detained by a state party abroad is thereby within its authority and control, and thus jurisdiction. It rather on both occasions cites to para. 175 of the judgment, where it had concluded that Russia had exercised spatial jurisdiction after 12 August 2008. Why does it opt for such an oblique reference to a jurisdictional basis that on its own terms cannot cover all of the detainees? It’s likely that the Court’s contortions are a product of its desire to avoid using explicitly the personal conception of jurisdiction, because doing so would have emphasized the arbitrariness of the position that detaining people is authority and control but killing them is not. Similarly, employing the personal conception of jurisdiction, which requires the exercise of authority or control by a state agent , would have required the Court to directly say whether the conduct of the South Ossetians detaining the victims was attributable to Russia. So there you go – mostly, inter alia, I suppose.

Jurisdiction regarding investigations into unlawful deaths

Just one more jurisdictional thing – I know this is getting a bit tedious! Recall how the Court did find that Russia had to investigate any allegations of unlawful killing even during the active phase of the conflict. Here’s how it did that:

  1. In the present case, in view of the allegations that it had committed war crimes during the active phase of the hostilities, the Russian Federation had an obligation to investigate the events in issue, in accordance with the relevant rules of international humanitarian law (see paragraphs 323-324 above) and domestic law (see paragraphs 48-53 of the admissibility decision). Indeed, the prosecuting authorities of the Russian Federation took steps to investigate those allegations (see paragraph 317 above). Furthermore, although the events which occurred during the active phase of the hostilities did not fall within the jurisdiction of the Russian Federation (see paragraph 144 above), it established “effective control” over the territories in question shortly afterwards (see paragraph 175 above). Lastly, given that all the potential suspects among the Russian service personnel were located either in the Russian Federation or in territories under the control of the Russian Federation, Georgia was prevented from carrying out an adequate and effective investigation into the allegations.
  2. Accordingly, having regard to the “special features” of the case, the Court considers that the Russian Federation’s jurisdiction within the meaning of Article 1 of the Convention is established in respect of this complaint (see, mutatis mutandis, Güzelyurtlu and Others, cited above, §§ 191-97). It therefore dismisses the preliminary objection raised by the respondent Government in this regard.

Let’s say this is right (my own view is simpler and broader – that the procedural obligation to investigate should apply whenever the state is faced with a credible allegation that one of its own organs or agents violated Article 2). The problem with this kind of multi-factorial, ‘special features’ analysis is precisely that there is no way of knowing how important any of these factors is individually, and how they would translate to other contexts. For example, if the Russian authorities took no steps whatsoever to investigate the allegations, would that have mattered? If Russian authorities had withdrawn on 13 August 2008, would that have mattered? If some of the suspects were not on Russian territory any longer, would that have mattered? How precisely does this apply to Eastern Ukraine, Nagorno Karabakh, drone strikes without territorial control? And so forth. There’s just no way to be sure which ‘special features’ tip the balance for the Court – but this holding did, to some extent, take the sting out of the no-jurisdiction holding on the substantive aspect of Article 2. And government and military lawyers should pay special attention to this holding, because the Court has effectively said that the duty to investigate possible war crimes will apply even if the war crimes themselves were not committed within the state’s jurisdiction.

Separate opinions

This is already an exceptionally lengthy post, and so I cannot here engage properly with some of the very rich and detailed 9 (!!!) separate opinions in the case. So let me just give you a quick punch line for each of the opinions, which will suffice to illustrate the tensions between the majority and the minority. Their main differences are on the Article 1 jurisdiction issue in the active hostilities phase – otherwise the Court is either unanimous or very nearly-unanimous on everything else. Some of the opinions also explore the issue of how the Convention would interact with IHL if it did apply to uses of force during active hostilities, which the Court avoided by dealing only with those questions (such as ill-treatment) for which no possible conflict between the two bodies of law arose.

Judge Keller defends the reasoning of the majority, describing the minority views as ‘ultimately founded on an overly expansive vision of the Court as an adjudicator of the totality of armed conflict.’ She, however, devotes a significant portion of her analysis to how the reference to ‘lawful acts of war’ in the Art. 15 derogation clause should be interpreted in some other case, in which the jurisdictional threshold was satisfied.

In his opinion, Judge Serghides argues tentatively for the solution that I have long advocated, which is that negative obligations to respect human rights should not be subject to any territorial or jurisdictional limitation. He nonetheless voted with the majority – note, not the minority – because the case was pleaded exclusively under Article 1 of the Convention, and in his view the negative obligation theory must be based in other substantive provisions of the ECHR. In short, had Georgia reframed its argument somewhat the active hostilities vote could have been 10 to 7, rather than 11 to 6.

Judge Lemmens, in a brief but effective opinion, laments the Court’s retrograde steps. He would have effectively overruled Bankovic, and finds much of the Court’s line-drawing to be arbitrary.

As for Judge Grozev, he takes a very different tack and would have applied a variant of the espace juridique/vacuum of protection approach, whereby the Convention would have applied to hostilities in this case but only because Georgia was already a party to the ECHR – it would not have applied if Russia had invaded some non-ECHR state. This is a troubling position, echoing the House of Lords in Al-Skeini but incongruent with the Grand Chamber judgment in that case. There seems to be no good reason of principle why an invasion of Georgia or Ukraine would properly be a human rights issue, but an invasion of say Belarus or Iraq would not.

In their joint opinion, Judges Yudkivska, Pinto and Chanturia extensively discuss the interplay between the ECHR and IHL if the Convention did apply. Their main argument is that Art. 15 derogations are necessary to accommodate IHL in those situations in which it is possibly in norm conflict with the Convention, especially with regard to deprivations of life and liberty. In another joint opinion, Judges Yudkivska, Wojtyczek and Chanturia argue for a flexible interpretation of the personal conception of Article 1 jurisdiction, and would depart from the rigidity of the Bankovic approach. Judge Pinto, writing separately, similarly laments the revival of Bankovic, arguing that the Court ‘will face a gargantuan task to restore the damage to its credibility caused by this judgment.’

Judge Dedov endorses the approach of the majority, while arguing (somewhat inconsistently) that Bankovic itself was wrongly decided on its facts. He spends most of his opinion on the political context of the case, which he argues led to a lot of uncertainty in the evidence. Even so, recall that he voted in favour of many of the findings of violation in the case. Finally, Judge Chanturia launches an all-out assault on the majority’s reasoning regarding the active hostilities phase of the conflict and Article 1 jurisdiction, arguing that it creates a vacuum in the protection granted by the Convention.


So, what should be our general take on this judgment?

It is Bankovic redux in more ways than one. I doubt that the judgment’s reasoning, with all its ambiguities and contradictions, will find many defenders in the academia, although we will all keep parsing its paragraphs. It will definitely find a warmer reception among government and military lawyers, but even there its caveats, open doors and sheer lack of clarity will make it difficult to digest – note, in particular, the Court’s broad but vague approach to the applicability of the procedural duty to investigate.

Take also, as another example, the fact that (as in Hassan) the Court expressly confined its holding to IACs. Where does that leave the more common NIACs and uses of force short of armed conflict? While in Hassan there were some relevant structural differences in IHL as it applied to IACs and NIACs (e.g. as to the authority to detain) which could justify confining the Court’s holding to IACs only, that does not appear to be the case here. And there is simply no legal or logical reason why the inquiry into state jurisdiction under Article 1 should depend on the existence of an armed conflict, let alone on its classification.

Or, consider a scenario in which the Court was faced with claims of violations by both parties to an IAC, which is precisely the posture in the recently filed Armenia-Azerbaijan cases re Nagorno Karabakh. (In other words, imagine if Russia had also sued Georgia). Will the Court find that both states lack jurisdiction in respect of ongoing hostilities, or that only one of them (the non-sovereign) lacks it? And if so, how could it possibly justify such disparate treatment of the two warring parties?

Similarly, what of kinetic uses of force that are not artillery shelling and aerial bombardment, but are more ‘proximate’ or ‘specific’ (whatever those words mean exactly)? Hand to hand, gun to gun, tank to tank, drone to target? Would the personal notion of jurisdiction capture any such scenarios? Note in that regard that the Court did not rely on the ‘public powers’ concept which it had imported into the personal conception of jurisdiction in Al-Skeini as a limiting factor. What are we to make of that? And what are we to make of those cases – probably the majority of cases – in which the line between active hostilities and some kind of stable control is much blurrier? Or of extraterritorial assassinations in peacetime?

The judgment is, in sum, precisely the kind of exercise in arbitrary line-drawing that the Court has engaged in before, and precisely the kind of exercise which ultimately always proves unsustainable in the long run. The big question is how long that long run turns out to be. On a very minimalist reading, all the Court did here was to preserve the outcome Bankovic – as Al-Skeini did too – but no more than that. However, its furtive treatment of detention, without direct reference to Hassan and the personal conception of jurisdiction, is a more problematic signal. And in any event it is reasonably clear that this majority of the Court wanted to strongly discourage the litigation of armed conflict issues before it. That is the basic message that the Court is sending. In terms of its immediate implications, the judgment will definitely make it more difficult to pursue any of the pending interstate and individual cases with an ‘active hostilities’ (Context of Chaos) component, e.g. those regarding Eastern Ukraine or Nagorno Karabakh in particular. As for the reasonably broad holding on the applicability of the procedural duty to investigate, we will see its implications shortly in the forthcoming judgment in Hanan v. Germany, dealing with the Kunduz incident in Afghanistan, which will likely be handed down soonish. 

That said, the reliability of any such predictions is undermined by the amount of time the Court takes to process these cases (13 years from application to the merits judgment in this one) and the shifting composition of the bench in the meantime. Only a couple of the judges in Bankovic Grand Chamber sat in the Al-Skeini Grand Chamber; none of the Al-Skeini judges sat in the Georgia v. Russia No. 2 Grand Chamber. The terms of office of many of the judges in this Grand Chamber (in the majority and minority both) have already expired or are about to expire. And any Grand Chamber is in any event only a partial representation of the 47 judges of the Court. With all the division demonstrated in the separate opinions there is simply no way to be sure how future Grand Chambers (or mere Chambers) will look at Georgia v. Russia No. 2. The judgment certainly goes against the expansive general trend in international human rights law regarding both extraterritoriality and the applicability of human rights in armed conflict, but it is premature at this stage to say with any confidence whether the trend is being reversed in Strasbourg for the foreseeable future, or whether this judgment will only be a momentary blip, albeit a very large, contextually chaotic one. Time will tell.

Disclosure: In 2014, I was briefly retained by Georgia as part of its legal team in the written pleadings stage of the case, but I have not had any involvement with the case since. To be clear, none of my comments above are based on any inside knowledge.

Print Friendly, PDF & Email

Leave a Comment

Comments for this post are closed


Ed Bates says

January 25, 2021

We are in debt to you for such a fantastic post, distilling the judgment in such a readable way - thank you!

Re Judge Grozev: yes, I see your point that, 'There seems to be no good reason of principle why an invasion of Georgia or Ukraine would properly be a human rights issue, but an invasion of say Belarus or Iraq would not'.

However, does Judge Grozev not approach the matter from a slightly different angle? The reason he gives is a little less pure ‘human rights’ and a little more going to the specific identity/ nature and qualities of the special legal regime that the Convention is *as a human rights instrument*... and so the relationships/ obligations it creates/ establishes for its *member States*.
[Judge Grozev: "First, it [the Court] has underlined that the Convention is a regional international treaty ensuring the protection and collective enforcement of the rights enshrined in the Convention in an essentially regional context, and notably in the legal space of the Contracting States. Second, taking into account the purpose and objective of the Convention as a human rights treaty, it has insisted on guarantees for the effective protection of human rights. It has held that no black holes can exist in the system of human rights protection in the legal space under the Convention. To allow this would be entirely at variance with the fundamental principles of the Convention"].

Personally, I see strength in Judge Grozev’s position – justifying the line-drawing by the nature of the Convention and the legal obligations is creates between member States in the field of human rights (cf his last line: ‘While a State’s jurisdictional competence for the purposes of Article 1 of the Convention is primarily territorial, exceptional circumstances do exist, on the basis of the mutual, bilateral undertakings of the Contracting Parties to guarantee effectively the rights set forth in the Convention, where a Contracting State engages in acts of war on its own territory or on the territory of another High Contracting Party to the Convention’).
Of course, this would still leave disappointment for many for IACs outside of ECHR space. I would see that as fitting with your point re ‘we’re oh so very sorry (‘may seem unsatisfactory to the alleged victims’), but war [outside ECHR space] is not our problem. Let others try to solve it – states, IHL, the UN, international criminal courts – but not us’. [The reasons being, the ECHR is specialised as set out above]

Alexander Bedrosyan says

January 25, 2021

What a comprehensive and insightful summary. Thank you, Marko.

Marko Milanovic says

January 26, 2021

Many thanks for the comments.

Ed, thanks in particular for raising this issue. To be honest I think that this is a distinction without a difference. It is certainly true that the ECHR is a regional treaty, and that each state party assumes obligations towards other states parties and not towards third states. That means, for example, that say Serbia could sue Russia for HR violations in Russia or Georgia, but that say Belarus can't sue Russia regardless of where the violation happened.

But from that entirely uncontroversial proposition one simply cannot deduce the conclusion that the ECHR (in armed conflict) only applies to conflicts that take place on the aggregated territories of its states parties, but not elsewhere. A logically doesn't entail B, as it were. Nor does B follow from any provision in the Convention, any rule of interpretation or broader principle. So the only way that particular position should be regarded as 'correct' is if it was somehow normatively desirable, as measured against the values enshrined in this legal regime. And I just can't see how it can be so justified, from the standpoint of UNIVERSAL human rights protected by a REGIONAL treaty. In other words, I just don't see how one could argue that the position that the ECHR would apply if say the UK bombed Paris but not if it bombed Minsk can cohere with the idea that INDIVIDUALS, not just states, possess rights inherent in their dignity. Yes, it is true that if (say) the Royal Air Force bombed Minsk the state of Belarus could not invoke the ECHR against the UK, but it is not true that individuals affected by the bombing could not do so themselves.

John says

January 26, 2021

Dear Marko,
Many thanks for this analysis. Do you have any thoughts on the following:

1) Does it matter if you generate your own chaos for the purpose of evading extraterritorial jurisdiction (ETJ)?

2) If ETJ has occasionally been rejected for uses of force on the basis that they were merely "instantaneous" (para 124) is it fair to then attempt to distinguish other historic cases, in which ETJ was found, partly on the basis that the force was only "isolated and specific"(para 132)?

3) Is the added "element of proximity" in the latter cases a conflation of the spatial and personal models of ETJ? Could you personalize an otherwise insufficient spatial claim to satisfy article 1?

4) Is the proximity element further invitation for states to avoid boots on the ground and to rely increasingly on less discriminate aerial bombardment?

5) Wasn't state agent authority and control also capable of being made out by the assumption of public powers, viz. security operations, in otherwise equally chaotic NIACs (see House of Lords comments in Al Skeini)?

6) Are the more detailed rules of IHL in IACs (c.f. NIACs) influencing the decision on jurisdiction?

7) With this and Hassan do you think states will be more inclined to adopt the Akande analysis that intervention in external NIACs without host nation consent would also amount to an IAC? What if it is a NIAC and IAC?

8)Assuming the same degree of chaos in the border regions, would attacks against ISIS in Syria be considered outside the UK's jurisdiction (on the Akande IAC analysis) whilst attacks against ISIS in Iraq (with host nation consent) would be inside the UK's jurisdiction?

9) If the Convention is to be interpreted in harmony with other rules of international rule why is there no reference to the jus ad bellum?

10) Would a Russian friendy-fire incident within the active phase of hostilities also have been outside jurisdiction?

11) How do you like the floodgates argument at para 141?

Marko Milanovic says

January 27, 2021

Dear John,

Wow, that's quite a list. Let me take a stab.

1) Depends on the chaos I guess. If we're talking strictly about the question of control over territory as a basis for jurisdiction my view would be that it shouldn't matter whether the territorial state lost control through a 'fault' of its own - this would normally be a fraught political question, e.g. with Georgia/Russia. Nor should it matter whether the intervening state is acting lawfully or unlawfully. What matters is the fact of control.
2) Completely agree. The whole instantaneous, isolated, specific, proximate stuff is just arbitrary.
3) Can you give me a scenario here?
4) Yes. Perverse incentives. As with Bankovic, which incentivizes (at least nominally, hard to tell how much practically) killing versus capture, air versus ground operations.
5) No. I don't think the idea of public powers is properly understood as a separate basis for jurisdiction. The term comes from Bankovic, where it is used with regard to the spatial model. Then it makes a re-appearance in Al-Skeini (Grand Chamber), but by reference to the personal model, effectively as a way of (arbitrarily) limiting it. See my EJIL piece.
6) Probably. More likely it's that it is somehow more politically palatable for the Court to exempt active IACs from scrutiny. In any event with regard to say killing it makes little sense to distinguish on the basis of the nature of the conflict, since the rules are basically the same in both types of conflict.
7) Maybe, I don't know. Doubt it will have much effect. It would really depend on how durable this is as a precedent, and what the Court ultimately decides on NIACs.
8) That would be a completely arbitrary result.
9) That's a perfectly valid question. Schabas e.g. has argued that ad bellum illegality should always entail a violation of the right to life. That's the position also of the HRC in GC 36. My own view is that we should probably erect a wall here as with as bellum/IHL. But both positions are plausible.
10) Another great question. There have been some similar cases before the UK courts, albeit dealing with bad equipment issues (Smith 1 and 2). The ECtHR had one such case, Pritchard, but the UK settled it after Smith 2. Personally I don't see how it could be justified normatively to say that if UK soldier A by mistake triggers a grenade killing UK soldier B and Iraqi civilian C, B would be within the UK's jurisdiction but C would not.
11)As I noted in my post, I think that's what determines the whole case. I neither like it nor dislike it - the problems are real, so is the Court's fear. The issue is what to do about it, and my view is that such questions can and should properly be addressed on the merits (e.g. on issues of evidence), rather than through a jurisdictional analysis.

John says

January 27, 2021

Dear Marko,
Thanks so much for your excellent replies. Really appreciate it.

John says

January 30, 2021


On the jurisdiction point, if State A is invaded by State B is State A obliged to honour its ECHR obligations (subject to any derogation) whilst aggressor State B is not (and needs not even derogate unless State A pushes its forces back into State B)?

On Keller's observations about derogation, if article 15(2) envisages derogation for "deaths resulting from lawful acts of war" doesn't that imply that such deaths are capable of being within jurisdiction in the first place? Given the context in which the Convention was written, surely the drafters must have envisaged soldiers losing lives in their millions by fighting "extraterritorially" in flanders fields etc? I fail to see that they meant to confine this provision to when effective control of an area had been established extraterritorially or when they were taking POWs (not least because you can't lawfully kill POWs and the high threshold for ECA suggests your active hostilities are more likely to extend beyond that ECA).

DH says

January 30, 2021

Dear Marko, thank you very much for this thought provoking and informative piece.

Just out of curiosity, I was wondering what is your take on the impact of the decision of the Court with regards to the active hostilities phase on the much awaited judgment of the Court in the case of Hanan v. Germany, where the applicant argued that the airstrike that was taken in the Kunduz region fell under the jurisdiction of Germany? Many thanks.

Marko Milanovic says

January 31, 2021

Thanks again for the comments.


As for your first question, that's the position as I explained towards the end of the post in Armenia/Azerbaijan, or if Russia had also sued Georgia. The Court left this issue open. Its Ilascu jurisprudence could easily lead to such a result. Personally I think that would be hard to justify, i.e. imposing disparate obligations on two parties to an IAC would be arbitrary. But that's a possible reading of the Court's position at the moment.

As for your second question, the text of Art. 15(2) unequivocally demonstrates that the drafters intended the Convention to apply in 'war'. 'War' at a minimum must include contemporary IAC. But we have no idea from the drafting history whether the drafters thought that this would include any extraterritorial component of the IAC, as in the example you give of British soldiers fighting in Flanders. (See my piece on extraterritorial derogations for a more extensive analysis). Whatever the right answer to this question is (and to my mind it must be in the affirmative), we must get to it from first principles rather than on the basis of the intentions of the drafters.


I have no way to be sure. But if the Court found the obligation to investigate to apply even during the active hostilities phase of Georgia/Russia, it would really have to contort itself to find otherwise with regard to Afghanistan. But they are very good at self-contorting. So I have no way of knowing what they'll decide in Hanan, which again I imagine will be handed down quite soon.

Onisime says

February 11, 2021

Thank you for the wide-ranging and interesting analysis.
You made an excellent comment that imposing disparate obligations on two parties to an IAC would be arbitrary. Do you think that the criteria for determining extraterritorial jurisdiction of a state and the criteria for determining the ability of a territorial state to exercise jurisdiction over disputed territories would be the same while the Court deals with individual applications against Georgia covering the period before 12 August? If there was no actual control over area or individuals during the chaotic phase of IAC, and extraterritorial jurisdiction was excluded based on this argument, then why not to argue that the same criteria (the ones applied to extraterritorial jurisdiction) can be applied to the territorial state that was unable to re-establish control over the disputed area outside of its authority for approximately twenty years? Is it possible to distinguish between positive and negative obligations of the territorial state and argue that during the active phase of hostilities (IAC), chaos and uncertainty related to territorial control, a territorial state is only bound by positive obligations?

Raphaël van Steenberghe says

March 19, 2021

Thanks for this very interesting post.
Just to give information about this strange §276, in which the Court referred to a "strict control" exercised by Russia over the South Ossetian forces.
The French version did and does not contain that sentence. The English version has been modified on 29 January 2021 (maybe thanks to your post!) and the sentence was deleted (cf. footnote 32). Your guess might have been right!
Raphaël van Steenberghe