The ICJ’s Treatment of Questions of Occupation in Gaza

In its recent Advisory Opinion, the ICJ highlighted that “Israel’s withdrawal from the Gaza Strip [had] not entirely released it of its obligations under the law of occupation”, specifically because, having “established its authority”, it “remained capable of exercising, and continued to exercise, certain key elements of authority over the Gaza Strip” (para 93).

Milanovic has highlighted that the Court’s ruling was “ambiguous”. Specifically, it is unclear whether the Court intended to describe Israel as having been an Occupying Power in Gaza post-2005, or whether it intended to delineate certain post-occupation obligations (as Judge Iwasawa considered in his Separate Opinion at para 8). The Court’s findings are perhaps best read as suggesting that Israel remained occupier post-2005. Such a reading is most compatible with its treatment of the Occupied Palestinian Territory (OPT) as a single entity throughout its Advisory Opinion in its assessment of Israel’s policies and practices. The Court highlighted this, holding that – “from a legal standpoint, the Occupied Palestinian Territory constitutes a single territorial unit, the unity, contiguity and integrity of which are to be preserved and respected” (para 78). Such a view is also perhaps most compatible with the Court’s holding that Israel’s obligations would remain commensurate with its degree of effective control, which appears to imply a continuing state of some form of occupation (para. 94).

The ICJ’s treatment of the state of occupation in Gaza

However, the Court’s approach in its Advisory Opinion is questionable. While it rightly accepted the functional approach to occupation, I doubt whether Israel was indeed capable of exercising its authority in Gaza sufficiently for its occupation to be found as having continued post-2005 (and before the current Israeli military operation, which was temporally outside the Court’s purview). Furthermore, in my view, the practices the Court relied on to find that Israel had indeed exercised elements of authority in Gaza – including restrictions on the movement of peoples and goods, the blockading of the Gaza Strip and the military buffer zone – are, as purely external methods of control, insufficient to constitute the “[exercise of] key elements of authority over the Gaza Strip” (para 93). Instead, the Court should have relied on Israel’s continued exercise of administrative authority vis-a-vis Gaza residents to find the existence of a state of occupation.

The Court appeared to accept a “functional” approach, as initially developed by academics like Scobbie and Gross (particularly in the latter’s book The Writing on the Wall), to the continuation of occupation in its Advisory Opinion by relying on the ability to exercise authority rather than the actual exercise of authority over a territory. Gross’ view had already found favor with a variety of States, academics (such as Jaber and Bantekas) and international organizations, including the ICRC. However, support for the “functional” approach was not universal – Shany has suggested that Israel cannot be an occupier without the physical presence of its troops but may still have obligations under human rights law and the jus in bello, while states like Israel and the U.S. have rejected the claim that the former remained an occupier in Gaza post-2005.

The Court moved away from its previous traditional approach in Armed Activities, where they relied on the actual substitution of authority (via the Ugandan appointment of a governor in Ituri province). This is easily justifiable because Armed Activities was concerned with when an occupation began, not ended. Either way, this approach is welcome – as Ferraro wrote, the previous test in Armed Activities was illogical, and could have potentially led to a finding that Germany did not occupy Denmark because it “allowed the Danish government to function, despite its military supremacy”.

The Court should not have found that Israel remained capable of exercising its authority within Gaza. Firstly, the standard for whether Israel “remained capable of exercising its authority” would rely on whether, per the ICRC (at page 12), Israel could “reassert its full authority in a reasonably short period of time” (emphasis added). The Court appeared to accept this position in its Advisory Opinion, holding that physical military presence was not required for a continuing state of occupation. Hamas launched a large-scale attack on 7 October 2023, killing over 1,100 people in Israel, including more than 700 civilians. Despite a military response that has been often criticized in scale, including by Israeli allies, Israel has still, almost a year later, been unable to re-establish full control over the Gaza Strip, with a significant Hamas presence remaining and fighting continuing. Even if full control was not necessarily required for an occupation to continue, Hamas retains significant administrative authority over Gaza, including running its Ministry of Health. These facts suggest that Israel, particularly after Hamas established its administration of Gaza, has not been sufficiently capable of exercising its authority within the Gaza Strip, at least before Israel’s recent military operation.

Second, the practices that the Court relied on were insufficient to suggest that Israel did indeed exercise its authority within Gaza, but the Court could have found the existence of a state of occupation with reference to other Israeli practices. The test (under the functional approach) that the ICRC supported in 2015 relied on the “exercise [of], within all or part of the territory, governmental functions acquired when the occupation was undoubtedly established and ongoing” (emphasis added). While the usage of coercive measures to control the population of the Gaza Strip from outside undoubtedly has a significant humanitarian impact on Gaza residents, it would not be coherent as a matter of law to suggest that those practices alone would be sufficient for a state of occupation to continue. Rather, such “external” control measures are better understood generally as being part of “sieges” or “blockades”, although the existence of a siege/blockade is admittedly not necessarily incompatible with a state of occupation.

This is best seen from the commonalities between the Israeli practices relied upon by the ICJ and other situations more clearly recognized as siege operations. For instance, Syrian rebels (while unable to establish air superiority) sieged Syrian government forces in Nubl and al-Zahraa for more than three years, with residents having significantly restricted access to food and petrol, among other basic necessities. Similar restrictions were seen in the Siege of Sarajevo, where the siege of the city by the Serbians left Bosnians without sufficient basic necessities and largely without the ability to freely move in and out of Sarajevo. While most such “siege” or “blockade” situations have never been adjudicated before an international court, the Geneva Conventions (GC) and their Additional Protocols (AP) appear to envision such situations as different to occupations – as seen by the distinct protections offered at times to “blockade” situations, including the requirement generally to allow relief according to Article 70 AP I, although this is also subject to agreement by the combatants (Art. 70(1)). Of course, it must also be acknowledged that Israel’s degree of control in this regard is more significant than in most other enclosure situations – as seen by its construction of a border wall around Gaza, the first iteration of which was as early as 1994, when Israel was undeniably an occupier of Gaza.

The similarities between the practices the Court relied on, and other situations of enclosure, suggest that the practices the Court relied on would best be treated (on the basis of lex lata) as forming a siege/blockade situation, rather than an occupation. As I have acknowledged, the elements constituting enclosure and occupation situations will often overlap, and both share the crucial commonality of a high degree of coercive control imposed on a local population. However, the presence of such practices alone cannot be sufficient for a state of occupation to exist without more. Such a view would unjustifiably lower the threshold for occupation and bring the law of occupation closer to the discredited Pictet theory, which, as Bothe summarizes (at page 38), suggests that “any successful invasion creates a situation of occupation”. The Pictet theory has been criticized for multiple reasons – notably including the unrealistic obligations it would impose on States, which would discourage compliance with international humanitarian law (IHL).

The Court should have still found that Israel was an Occupying Power in Gaza, however. Israel continued exercising sufficient administrative authority in Gaza for the finding of a state of occupation, particularly through its continued administrative control over the Palestinian Population Registry (including in Gaza), which records Palestinian demographic information, both in the West Bank and the Gaza Strip, and enables a significant amount of Israeli control over the Gaza Strip. Furthermore, Israel continued controlling some taxation destined for Gaza, purportedly for transfer to the Palestinian Authority, but those funds were temporarily frozen after October 7. These manifestations of authority are themselves sufficient to find a continued state of occupation. If that is insufficient, Milanovic wrote in 2009 that it might be possible to find some positive obligations owed to Gaza by Israel, perhaps as a result of reparational duties owed by Israel to Gaza as a result of the occupation (to which the Court agreed in its Opinion). Such post-occupation obligations could also be derived from an expansive understanding of Israel’s duties as usufructuary under Article 55 of the Hague Convention (IV) of 1907, which would require Israel to not deplete the OPT of resources while also maintaining public buildings, real estate and certain other facilities. Israel, as usufructuary, could have a duty to replenish that which it had depleted or otherwise damaged under its occupation.

A lacuna in protection?

However, the difficulties posed by the unique occupation in Gaza, partially enforced through external control methods, expose a different problem – a lacuna in the protections available to civilians between the conflict/invasion stage and the occupation stage. There is a cliff-edge of protections between conflict and occupation situations – when a conflict becomes an occupation, civilians enjoy far more protections from the Occupying Power than they do when the state is merely conducting military operations within the territory. Alas – the protections of occupation do not apply in sieges/blockades (that are not also occupations), meaning that civilians do not enjoy the elevated humanitarian protections of occupation despite the significant and unique challenges that they face in such “enclosure” situations, including entrapment, displacement and more.

International law only provides few protections that are relevant to civilians in a “siege” stage. States are obligated not to use starvation as a method of warfare (Article 54(1) AP I), although a violation of that obligation would require deliberate intent (Commentaries to AP I, para 2089), which would often be difficult to prove in scenarios of warfare. There is also a stronger obligation under Article 70 of AP I to “allow and facilitate rapid passage of all relief consignments” in situations of conflict, although that would also be dependent on the agreement of the parties involved, and is certainly a weaker protection than the obligation under the law of occupation, where under Article 55 GC IV the occupier must even, unless impossible, “bring in the necessary foodstuffs, medical stores and other articles if the resources of the occupied territory are inadequate”. These protections do not go far enough, and can lead to insufficient humanitarian protection in “siege” situations.

The ICJ’s proposed solution – imposing a “sliding scale” of obligations depending on the degree of effective control in an occupation – does not adequately protect civilians in “enclosure” situations, as an initial establishment of authority leading to the finding of a state of occupation would still have to be found for those protections to apply. Such a solution still deprives protections for civilians in situations where there was never an occupation, such as the abovementioned Siege in Nubl and al-Zahraa. Furthermore, the ICJ remained ambiguous, not going into any detail on precisely how the obligations owed would vary with the degree of effective control, including on whether there were any irreducible core obligations. This can reduce the certainty of the protections available in individual occupations, as Judge Cleveland appeared to endorse in her Separate Opinion (para 11), particularly given the varying elements of control between different occupations.

Instead, specific, stronger protections that go beyond weak existing protections must be adopted for “enclosure” situations. This would better protect civilians in such scenarios by moving away from the “cliff-edge” of protections under lex lata while providing states more certainty as to their IHL obligations. Moving away from the current bivalent distinction between occupation and invasion would also better reflect the large number of conflict scenarios that exist in reality. States must collaborate to adopt such protections with urgency, given the continuing frequency of “enclosure” situations in modern conflicts such as the Syrian Civil War and the Yugoslav Wars – and must endeavor to ensure a wide breadth of protections are indeed available to civilians.