In my previous post, I discussed the bottom line of the ICJ’s UNRWA advisory opinion – essentially the Court’s finding that Israel’s obligations under IHL, IHRL and the UN Charter compelled it not to obstruct UNRWA’s work, in the particular circumstances of the occupied territories. I did not discuss in that post the issue on which the judges seemed to have been the most divided, namely the interpretation of the duty to cooperate with the United Nations under Article 2(5) of the Charter, and I will not discuss it here either – Eliav Lieblich has an excellent post on Just Security that examines this question in detail.
In this post, I will cover some issues that I personally found to be very interesting. But the AO raises many others; I am not aiming to be comprehensive (but I will be a bit nerdy). First, I will discuss the Court’s approach to the status of Gaza as an occupied territory. Second, its application of human rights law. Third, how the Court conceptualized the inviolability of UN premises during armed conflict. Lastly, how the Court cited the ICRC Customary IHL Study.
Gaza as occupied territory, and Israel as the occupying power
Much of the UNRWA AO is premised on the applicability of the law of occupation, specifically the various provisions of the Fourth Geneva Convention. That is uncontroversial as to the West Bank, but has been a point of some contention as to Gaza. Broadly speaking, many international law scholars and essentially the whole humanitarian community regarded Gaza as occupied even after Israel’s 2005 disengagement, while many lawyers with a military background regarded Gaza as not being occupied due to Israel not having boots on the ground and thus not having effective control over the territory (obviously all of this relates to the situation before 7 October 2023). In the 2024 advisory opinion on the legality of the occupation, in which it rightly held that Israel’s continued presence in all of the occupied territories was illegal, the Court adopted a variant of the so-called functional approach, holding that Israel’s obligations with regard to Gaza (pre-7 October) varied with the degree of its effective control (2024 AO, para 92).
However, as I explained in an earlier post, the Court’s 2024 AO was carefully crafted to retain a certain degree of ambiguity as to the status of Gaza, presumably in order to enable consensus on the bench. In particular, while the Court held that the law of occupation applied to Gaza, it did not expressly say that Gaza, wholly or partly, was occupied. The AO was drafted so that it could also be read as holding that Gaza was indeed occupied. But it could also be read as holding that Gaza used to be occupied, that now it may or may not be occupied, but that Israel nonetheless retained some residual obligations even if Gaza was not occupied. This ambiguity was expressly noted in Judge Iwasawa’s separate opinion:
Thus, while the Court makes clear that Israel continues to be bound by certain obligations under the law of occupation, it does not take a position as to whether Gaza remained “occupied” within the meaning of the law of occupation after 2005. […] The situation in Gaza has drastically changed since 7 October 2023. However, events taking place after that date are beyond the temporal scope of the Court’s inquiry (para. 1 above).
This ambiguity (Gaza still occupied v. not occupied but obligations apply residually and functionally) is often missed in some of the discussion of the 2024 AO, but again it was quite deliberate. (Note also in that regard the brief reference to the AO in the freshly minted 2025 ICRC Commentary on GC IV, para 378, which to my mind also seems to be very carefully drafted, if I can put it that way).
This brings us to the 2025 UNRWA AO. Here, obviously, the major material difference compared to the 2024 AO is that the new case deals with the post-7 October 2023 period, in which Israel has had an extensive military presence within Gaza, with boots on the ground in much, but not all, of the territory. If one was of the view that Gaza was occupied even pre-7 October, then subsequent events simply strengthened Israel’s hold over the territory. Conversely, those who thought that Gaza was not occupied before may now think that it is occupied today, either wholly or partly.
Here, again, the Court deploys some strategic ambiguity. Before quoting from the Court it is worth noting, in that regard, that Judge Iwasawa is now President of the Court and that he was ex officio a member of the drafting committee of the 2025 AO, since he was in the majority. I will quote here the three key paragraphs and their heading in full:
1. Israel’s status as an occupying Power in the Gaza Strip
85. A small group of participants contested Israel’s status as an occupying Power in relation to the Gaza Strip. The Court recalls that it found in the Advisory Opinion on the Legal Consequences arising from the Policies and Practices of Israel in the Occupied Palestinian Territory, including East Jerusalem that, after the withdrawal of its military presence in 2005, “Israel remained capable of exercising, and continued to exercise, certain key elements of authority over the Gaza Strip”, and that Israel’s obligations under the law of occupation “have remained commensurate with the degree of its effective control over the Gaza Strip” (Advisory Opinion of 19 July 2024, paras. 93-94). These findings were based on the control exercised by Israel over the Gaza Strip prior to 7 October 2023, including control of the land, sea and air borders, restrictions on movement of people and goods, collection of import and export taxes, and military control over the buffer zone (ibid., para. 93; see paragraph 43 above).
86. Since that date, Israel’s effective control over the Gaza Strip has increased significantly, as evidenced, inter alia, by Israel’s increased military control in large portions of the territory and Israel’s blocking of aid between 2 March and 18 May 2025, which prevented all humanitarian aid from entering into the Gaza Strip (see paragraphs 70-72 above). Therefore, the Court finds that Israel’s obligations under the law of occupation have also increased significantly, commensurate with the increase in its effective control over the territory. Those obligations include the obligations under the law of occupation considered in this section.
87. The Court observes that the fact that hostilities are ongoing does not necessarily preclude the simultaneous application of the law of occupation. When hostilities take place in an occupied territory, the law of occupation applies alongside other rules of international humanitarian law relating to the conduct of hostilities, and the occupying Power must comply with both sets of rules. However, the intensity of the hostilities could affect the implementation of certain obligations under the law of occupation, and therefore the particular conduct required of the occupying Power.
This all seems right to me – but the ambiguous drafting here needs to be acknowledged. It is interesting how the Court, in a sub-section posing the question whether Israel is the occupying power in the Gaza Strip, does not clearly say that it is. This is sort of implicit in the very last sentence of para 87, but note how that whole paragraph is framed in the abstract, rather than as being directly applied to the situation in Gaza. The Court also does not say anywhere in the AO that Gaza is an occupied territory, in whole or in part – that language is just not used, except when the Court refers to Gaza as part of the capitalized entity ‘Occupied Palestinian Territory.’ However, throughout its subsequent analysis the Court refers to Israel as the occupying power (see, e.g., paras, 92, 94, 128, 141, 157, 176); many of those references clearly include Gaza, and not just the West Bank, within their scope.
In short, the UNRWA AO is more consistent with the ‘Gaza is occupied’ theory than with the one that certain duties are applying residually post-occupation. But the Court is again being somewhat coy – one could, for instance, imagine that some of the judges thought that Gaza was not occupied pre-7 October, but was re-occupied in the hostilities that followed, and that the ambiguity we have accommodated their views sub silentio.
Only one judge discusses this issue separately – Judge Gómez Robledo – and he was not pleased with the Court’s coyness, noting at para. 3 of his opinion that ‘the Court’s pronouncement [in paras. 85-7 AO] appears to convey a desire to avoid expressly concluding that the Gaza Strip is now under full-scale occupation’ and at para. 6 that ‘there is no justification for the ambiguity in the Court’s reasoning and that a more explicit assertion was warranted, since it would have reinforced the Court’s findings on Israel’s obligations in Gaza.’
Does all of this even matter? It perhaps matters less for the case of Gaza, and more for applying the law correctly in some future situation. It matters for distinguishing between those situations in which an occupation clearly previously existed and then there was a unilateral withdrawal, from those in which the establishment of the occupation itself may be done by means other than having troops on the ground in (all of) the territory concerned. And it may matter for the application of the duty of the occupying power to maintain public order in the occupied territory, per Art. 43 of the Hague Regulations. As Eliav notes in his post, saying that (all of) Gaza is occupied today could enable the argument that Israel has the right to resume hostilities against Hamas in order to prevent Hamas from engaging in extrajudicial executions and other forms of violence in the territory.
To conclude, the UNRWA AO is of great value for the vexing questions of assessing the beginning and end of occupations, and of the beginning and end of the application of the law of occupation. But some of the ambiguity used by the Court simply means that it can’t be the final word on the matter, especially if we are thinking of scenarios beyond Gaza.
Human rights law in the UNRWA AO
IHL was definitely centre-stage in the AO, but human rights law also played an important role. The bottom line result was that Israel could not comply with its obligations under human rights law towards the population of the occupied territory if it impeded the work of UNRWA and other humanitarian organizations. In the process of getting there, the Court made several noteworthy points.
First, the Court reaffirmed its finding in the 2004 Wall AO and the 2024 Legality AO that Israel’s human rights obligations apply extraterritorially, to those territories that it does not own but is occupying (paras 148-9).
Second, in doing so, the Court held (para 150) that ‘the principle that a State’s human rights obligations extend to acts taken by that State in the exercise of its jurisdiction outside its own territory, particularly in occupied territories, applies also with respect to CEDAW, the CAT and the CRPD.’
This paragraph is important for at least two reasons. The Court is clearly saying that the ‘exercise of its jurisdiction’ is a concept wider in application than simply control over territory (particularly in occupied territories, but not limited to them). The Court thus opens the door to non-spatial conceptions of state jurisdiction in the human rights context of the kind already used by various human rights bodies, e.g. the personal model of jurisdiction as control over the victim by a state agent – think, for instance, of US drone strikes, amounting to extrajudicial executions, against suspected drug traffickers on the high seas. Moreover, the Court is holding that state jurisdiction is the relevant legal principle for the extraterritorial application of treaties that do not mention this term (CEDAW and the CRPD). This is a non-obvious approach to take, but it makes sense as a policy matter to maintain consistency between different parts of the human rights regime.
Third, the Court twice referred to customary human rights law (paras 146 & 151). While it does not say so expressly, a reasonable implication is that it regarded all of the rights listed in para 151 as being customary in character.
Fourth, the Court also repeatedly referred to the state duty to respect, protect and fulfil human rights (paras 146, 152, 153, op para in 223). I may be wrong, but I think this is the first time that the Court endorsed this terminology and even used it in the operative paragraph of the opinion – which (again unless I’m not mistaken) originated in Henry Shue’s scholarship in the 1980s and was then mainstreamed by various human rights bodies. The mainstreaming is now complete!
Fifth, the Court has also cited, as persuasive authority, two general comments of the Human Rights Committee (GC 36 on the right to life) and the Committee on Economic, Social and Cultural Rights (GC 12 on the right to food), at AO paras 154 and 155 respectively.
Sixth, in doing so, the Court did not just strengthen the authority of the treaty bodies, it also endorsed their interpretations of the law on very important points. For example, the Court endorsed the Human Rights Committee’s view that the right to life is not confined simply to those situations that directly result in the deprivation of life, but has a much broader dimension. Thus, ‘[t]he duty to protect the right to life also requires States to take measures addressing widespread hunger and malnutrition as well as extreme poverty and homelessness. Further, where necessary, States must ensure that individuals have access without delay to essential goods and services such as food, water, shelter, health care, electricity and sanitation.’
Seventh, the Court at no point even mentions the lex specialis principle – it saw no conflict between IHL and IHRL at all, and simply applied IHRL on its own.
Finally, the Court referred to many rules of IHRL addressing issues on which IHL is largely silent (e.g. mental health, the rights of peoples with disabilities, the situation of pregnant women – paras 155-160). All in all, this well demonstrates how IHRL and IHL can apply jointly in conflict situations, including those of occupation. In all of these aspects, I would submit that the opinion makes a major contribution.
UN immunities in armed conflict
This brings us to the third topic I wish to discuss. Paras 180 et seq of the AO deal with the question of the privileges and immunities of the UN, including UNRWA as a subsidiary organ of the General Assembly. Here too the Court makes several interesting findings.
First, in para 184 the Court concludes, quite correctly, that ‘within the territory of Israel, the presence and activities of the United Nations and its entities are subject to the consent of Israel. However, in the occupied territory, over which Israel, as an occupying Power, enjoys no sovereignty, it is not entitled to decide unilaterally, with respect to the presence and activities of the United Nations in and in relation to the Occupied Palestinian Territory, in the same way as in its own territory.’
Second, in para 185 the Court establishes that Article 105 of the Charter and the General Convention are not limited in their territorial scope, and that ‘[i]n the context of an occupation, an occupying Power exercises jurisdiction and control over the occupied territory and thereby assumes an obligation to respect the privileges and immunities accorded to the United Nations under Article 105 of the Charter and the General Convention in the occupied territory. This obligation, by its nature, derives from a State’s membership of the United Nations and its status as a party to the General Convention.’
Third, in para 186 the Court holds that UN privileges and immunities also apply during armed conflict. None of these three points can, in my view, be contested. The real question is not whether, but how, UN privileges and immunities apply during occupation and armed conflict.
In that regard, perhaps the most important is the rule set out in Article II, Section 3, of the General Convention, which provides for the inviolability of UN premises and their immunity from any form of interference. And here the key question becomes how that rule, which is framed in absolute terms, should apply in situations of armed conflict.
The main scenario here, one that Israel alleges has happened with some frequency, is the use of UNRWA premises, such as UNRWA-run schools, by Hamas for military operations – for example, to launch missiles against Israel, or to store weapons or house fighters. Under the targeting rules of IHL, a civilian object that is used for such purposes would become a military objective and would lose protection from attack. And so the question here becomes whether the UN privileges and immunities would apply on top of these rules of IHL and restrict what would otherwise be a lawful attack against such an object, due to the categorical nature of the protection for inviolability. (For more background and discussion of these issues, see this Articles of War post by Ori Pomson).
Here the Court reasons as follows:
- The obligation to respect the inviolability of United Nations premises and the obligation not to interfere with United Nations property and assets must also be upheld in the context of armed conflict, as such inviolability and non-interference are essential to safeguarding the independent and effective performance of the functions of the Organization under all circumstances. The Court acknowledges that the context of an armed conflict raises challenges, especially in the situation of potential loss of control by the United Nations over certain of its premises. However, it is for the United Nations to determine whether a particular facility remains the premises of the United Nations. In the view of the Court, such a determination by the United Nations creates a presumption that may only be set aside for the most compelling reasons and is to be given the greatest weight by States (cf. Difference Relating to Immunity from Legal Process of a Special Rapporteur of the Commission on Human Rights, Advisory Opinion, I.C.J. Reports 1999 (I), p. 87, para. 61). The obligation to respect the inviolability of those facilities qualifying as United Nations premises must be observed by all parties to the hostilities, along with the obligation not to interfere with the property and assets of the Organization. Damage to or destruction of the premises and other property and assets of the United Nations as a result of military activities may amount to a violation of obligations under Article II, Section 3, of the General Convention. (emphasis added)
There are many moving parts to this paragraph, and many things left unsaid – I did not find it altogether convincing, or as providing sufficient guidance. First, the Court acknowledges that UN premises can be used by parties to a conflict for military purposes (the ‘challenges’ it refers to). A rule saying that UN premises can never lose immunity from attack, no matter how they are used, can easily incentivize their misuse by parties to the conflict. Second, the Court, on the one hand, says that it is for the UN to determine whether a particular facility should remain inviolable, but then, on the other hand, says that this is only a presumption that can be set aside for the most compelling reasons.
So, consider a scenario in which Hamas stores ammunition and other supplies within an UNRWA school. My sense from what the Court is saying is that in such a scenario Israel would have to attempt to liaise with the UN to determine what is going on in that school and to put an end to any misuse – it could not simply proceed with an attack, even if such an attack would be lawful under IHL (subject to any proportionality considerations). But then the Court suggests that the inviolability presumption can be set aside for the most compelling reasons. One would imagine that one such reason would be fire at Israeli soldiers coming directly from an UNRWA building – surely in that scenario Israel could not reasonably be expected to engage in a diplomatic dialogue with the UN before acting. (Obviously, the analysis here assumes that Israeli forces are acting in good faith, rather than pretextually relying on the supposed misuse of these facilities in order to justify attacking them. I leave that factual point aside – my inquiry is about the approach that should be taken in principle.)
The Court indicates that this is how it should be understood by saying in the very last sentence that damage to UN buildings may amount to violations of inviolability, but are not always such. If this is right, it is unclear on what precise conceptual basis that conclusion should be drawn. Is it that UN premises lose inviolability because they are no longer under UN control, despite the UN saying that they are? Or is there some kind of implied exception, with unclear parameters, perhaps relying on IHL (note, again, the absence of any reference to lex specialis)? Or that necessity or distress as circumstances precluding wrongfulness could do some work here? It is difficult to know, and the Court just doesn’t explain itself fully. And what are we to do with the scenario in which an UNRWA building is not made the object of an attack, but is damaged incidentally when an attack was directed against a military objective close to it, or the scenario in which an UNRWA building is damaged due to an honest and reasonable mistake, which arose even though the attacker took all feasible precautions in line with IHL?
Finally, it is worth noting that the Court is saying that all parties to hostilities must observe the inviolability of UN premises. But again it is unclear on what exact basis the Court is making this pronouncement. In international armed conflict, that position would be straightforward in those cases in which the conflict is being fought between two UN member states. But in a non-international conflict, or one with mixed characteristics, that is less clear. Why, exactly, is Hamas as a non-state actor, which is not a party to the UN Charter or the General Convention, bound to respect the immunity and inviolability rules in those treaties? Again, IHL straightforwardly applies to organized armed groups, but I am not sure that UN privileges and immunities do and on what basis. A treaty can bind non-state actors which have not consented to it, just like custom can – but I wonder whether these two treaties do so. (I should also note that Judge Cleveland deals with some of these issues in her separate opinion).
Citations to the ICRC Customary IHL Study
I will conclude this post by look at a relatively small point, but one which I found to be of particular interest: how the Court cites the ICRC Customary IHL Study. Readers may recall that a few years ago Sandy Sivakumaran and I published an article examining the increasing authority of the Study, in part by measuring citations to the Study from international and domestic courts (IRRC article here, dataset here).
In recent years, many courts have cited the Study, thereby contributing to its authority. This includes, for example, various major judgments of the Grand Chamber of the European Court of Human Rights dealing with armed conflicts, including the Ukraine and Netherlands v. Russia judgment from earlier this year. The ICJ has, however, historically been particularly selective with its citation practices. While ICJ judges have cited the Study in their individual opinions, the Court itself has not done so – until now.
In para 84 of the UNRWA AO, the Study is cited for the first time; this is again done, for different propositions but in the same format, at several places in the judgment (paras 91, 135-6, 139, 143). Here’s how that looks on two occasions:
- The Court emphasizes the fundamental importance of the principle of distinction under international humanitarian law. Under the principle of distinction, parties to a conflict must at all times distinguish between civilians and civilian objects, on the one hand, and combatants and military objectives, on the other. Military operations may only be directed against combatants, and they must not be directed against civilians (see ICRC, Customary International Humanitarian Law Study, Volume II: Practice, Chapter 1, Section A, relating to Rule 1. The Principle of Distinction between Civilians and Combatants). As a corollary to this principle, the principle of proportionality prohibits attacks that may be expected to cause incidental loss of civilian life, injury to civilians, damage to civilian objects, or a combination thereof, that would be excessive in relation to the concrete and direct military advantage anticipated (see ibid., Chapter 4, relating to Rule 14. Proportionality in Attack). The principle of precaution requires a party to a conflict to take constant care to spare civilians and civilian objects. Accordingly, in the conduct of military operations, all feasible precautions must be taken to avoid, or at least to minimize, incidental loss of civilian life, injury to civilians and damage to civilian objects (see ibid., Chapter 5, Section A, relating to Rule 15. Precautions in Attack).
…
- The principle that humanitarian relief personnel must be respected and protected forms part of customary international law (see ICRC, Customary International Humanitarian Law Study, Volume II: Practice, Chapter 8, Section A, relating to Rule 31. Safety of Humanitarian Relief Personnel). Under Article 71, paragraph 2, of Additional Protocol I, applicable to occupied territories by virtue of Article 69, paragraph 2, of the same Protocol, States have an obligation to respect and protect personnel participating in relief actions. Article 71, paragraph 3, clarifies that the activities of the relief personnel may be limited and their movements may be temporarily restricted only “in case of imperative military necessity”. These provisions reflect customary international law.
I have no problem whatsoever with the Court citing the Study – on the contrary. What I found very interesting here is the idiosyncratic way in which the Court chooses to do so. The perceptive reader may already have noticed that the Court is always citing Volume II of the Study, assembling the practice that the ICRC used to compile the Study, and not Volume I containing the rules – the rules are only cited indirectly, in that the practice is ‘relating’ to them.
The reason why this is so interesting is that hardly any other court citing the Study ever bothers with Volume II (i.e. the database) on practice. They just cite the rules in Volume I. Don’t get me wrong, they like the comfort of knowing the practice is there, but they almost never evaluate any of it.
I can only guess why the Court decided on this particular citation format – just like I can only guess that much time and pedantry were spent within the Court and its Registry until they arrived at a solution that everyone was comfortable with. The Court (or at least some judges) wanted to avoid the impression that they are holding that any given rule is customary simply on the ICRC’s say-so. They may also wanted to give the impression that they did their homework and looked at all of that practice and did some verification thereof (which of course they didn’t). The Court likes establishing custom simply by assertion, and the ICRC Study was convenient in making those assertions appear more rigorous. In doing so, however, and despite its peculiar citation format, the Court inevitably added to the Study’s authority.
