Ethiopian Prime Minister Abiy Ahmed sent the country’s National Defense Forces (ENDF) into the northern region of Tigray last November in response to what he claimed was an attack on a federal military base by the Tigray People’s Liberation Front (TPLF), a group Ethiopia later designated a terrorist organization. Others in the region argue that the TPLF is simply the largest party in the legitimately elected Regional Government of Tigray and that the latter responded defensively to Abiy’s action by constituting the Tigray Defense Forces (TDF) as its regional armed force.

In the months following the ENDF deployment, Tigray has been ravaged by an armed conflict that has also drawn in armed actors from neighboring Amhara state and the Eritrean Defense Forces (EDF), each apparently with Abiy’s blessing. Throughout the fighting, there have been repeated reports of grave abuses. After eight months of intense violence, the conflict has now taken an unexpected turn. In mid-June, the TDF/TPLF launched a powerful counter-offensive, decimating and evicting Ethiopian and allied forces from the region, retaking the Tigrayan capital Mekele, and prompting Abiy to announce a unilateral (and ultimately short-lived) cease-fire on June 28.

The trajectory of the conflict is now uncertain, but the brief standoff did not bring respite for the civilian population in the region. Federal forces appear to be exploiting their control over what enters Tigray in order to put the region under siege, exacerbating the severe food crisis, in which 400,000 are already in famine, with a further 1.8 million on the brink. At the end of June, a local official described people “falling like leaves.” In a phone conversation with colleagues at the World Peace Foundation on the 9th of July, a doctor from Mekele depicted the situation as follows:

Though there is no killing in the street by bullet, people are dying in other ways. They’re dying of lack of medication and also food … People in Tigray are now in the complete dark, with no access to electricity, no phone, with no medical access, with no food supply. Everything is depleted. If this siege continues for a week or two weeks more, you will see that people will die in just the whole [place] – especially those in IDP [camps] will die. I think, the international community, it’s time now to intervene, at the moment. Otherwise, we’ll see the catastrophe, which is coming in the next month, probably.

Abiy has both denied that there is hunger in Tigray and promised that the situation will improve with immediate access for humanitarian organizations. The former claim is simply contrary to fact. The latter warrants careful scrutiny, particularly given the region’s geopolitical context and the scale and urgency of the need.

It is clear that occasional or sporadic deliveries will not be enough to avert profound devastation. Tigray is encircled. Although “accessibility” has improved due to the greater freedom of humanitarians to move within the region, the entry points into Tigray remain blocked and stocks are rapidly depleting. Those with control over what goes in will determine its fate. To the north, the region is bordered by Eritrea, which has been accused of many of the war’s worst atrocities. To the south is Ethiopia’s Amhara state, which has also been engaged in the conflict against the TDF/TPLF and has been accused of destroying or obstructing key access routes into the region. To the west is Sudan, but that border remains under the control of pro-government forces. To the east is Ethiopia’s Afar region, which had been relatively untouched by the conflict, but which has recently been the site of a significant TDF/TPLF offensive and subsequent fighting with regional pro-government forces.

Additionally, in late June, Ethiopia closed the airspace over Tigray. In response to external pressure, the government later announced that it would allow flights that undergo inspections, and the World Food Programme has since reported its first humanitarian passenger flight into Tigray. However, it remains unclear whether permissions will now occur at anything close to the necessary speed and scale.

Both as a matter of specifying a legal focal point around which political and diplomatic pressure might be mobilized to reverse the deterioration of conditions in Tigray, and as a matter of clarifying the prospect of accountability, a key question is whether the siege could meet the legal criteria of the largely untested war crime of starvation of civilians as a method of warfare.

The importance of that specific criminal category is exemplified by the fact that encirclement deprivation has typically been ignored in jurisdictions lacking the starvation crime. Characteristic features of siege deprivation complicate the applicability of more established war crimes categories, such as the killing of protected persons or attacks on civilian persons or objects. Siege deprivation need not involve attacks or destruction; it often involves impeding objects that could serve both civilians and combatants; there is not necessarily direct engagement between the besieging forces and the affected civilians; and the chain of causation from conduct to ultimate harm is long and complex.

On its face, the crime of starvation of civilians as a method of warfare offers a more appropriately tailored basis for accountability in such contexts. At the international level, it has now been codified in the Statute of the International Criminal Court (ICC) for both international armed conflicts (IACs) (art. 8(2)(b)(xxv)) and, pursuant to a 2019 amendment, non-international armed conflicts (NIACs) (art. 8(2)(e)(xix)). It is also included for both IACs and NIACs in the 2014 Malabo Protocol in the African Union system (arts. 28D(b)(xxvi) and 28D(e)(xvi)).


Neither Ethiopia nor Eritrea is party to the ICC and neither has signed or ratified the Malabo Protocol, which has anyway yet to enter into force. Nonetheless, the codification in these treaties reflects the widespread contemporary recognition of the crime, and there are various paths through which accountability might be pursued if it were found to be applicable to the siege of Tigray.

First, Ethiopia includes among the war crimes in its own criminal code, “wilful reduction to starvation,” destroying, removing, rendering useless, or appropriating “foodstuffs, crops, livestock, drinking water installations and supplies and irrigation works,” “hostile acts against or threats or insults to persons belonging to …. humanitarian relief organizations” such as the ICRC, and “any means or method of combat expressly forbidden” by the two Additional Protocols, which provide the underlying IHL prohibitions on starvation as a method of warfare (arts. 270(b, i), 276, 281). In the short term, these are relevant legal focal points around which to mobilize politically. In the longer term, they could provide the basis for domestic investigations and prosecutions. More controversially, it may even be that the current authorities in Tigray will endeavor to conduct their own trials of captured federal forces under this law, on the basis of their claim to be the legitimate regional government.

Second, a number of other states have also incorporated the starvation war crime (for both IACs and NIACs) into their domestic war crimes codes, many of which are underpinned by one or another variant of universal, or at least extended, jurisdiction (see, e.g., here and here (searchable database)). Among them is Germany (§ 11(1)(5)), which has been the subject of significant attention in this area for its use of a specialized war crimes investigative unit to pursue both Strukturermittlungsverfahren (structural investigations) and opportunistic individual investigations in responding to alleged war crimes in Syria.

Third, the codification of the crime in Ethiopian law and its likely customary status provide independent grounds for its straightforward incorporation into the statute of any hybrid or special tribunal that may be agreed and implemented as part of a future peace process (for comparison, see here, here, here, but see here).

Fourth, although ordinary territorial or nationality-based jurisdiction at the International Criminal Court is unavailable in Tigray (arts. 12(2)(a,b)), the fact of the crime’s codification in the Rome Statute is itself significant in providing a normative focal point for legal argument beyond The Hague. Moreover, the ICC remains a possibility in two scenarios. First the U.N. Security Council could refer a situation under article 13(b) of the Rome Statute, or Ethiopia could accept the Court’s jurisdiction on an ad hoc basis under article 12(3). Either of these can have retrospective effect, and there is no statute of limitations on the Statute’s core crimes. Although each of these paths to the ICC is politically unlikely in the short term (and the former in particular is open to a legitimacy critique), one need only look across the border to Sudan to see how the politics of engagement with the Court can change radically in a short space of time.

The Elements of the Crime

With accountability mechanisms available, if complicated, the key substantive question is whether the kind of encirclement deprivation threatening to broaden the famine in Tigray could qualify as criminal starvation. A challenge here is that, thus far, the crime has remained almost entirely dormant. It has yet to be invoked at the ICC and, for the most part, domestic jurisdictions have been equally inactive (for an exception, see Public Prosecutor v. M.P. et al.; on starvation jurisprudence more broadly, see the Starvation Jurisprudence Digest).

Against that vacuum, the elements of the crime provided in the ICC framework offer the most important interpretive starting point (Elements p.31, Resolution ICC-ASP/18/Res.5 annex II). Pursuant to that framework, the crime attaches when:

  1. The perpetrator deprived civilians of objects indispensable to their survival.
  2. The perpetrator intended to starve civilians as a method of warfare.
  3. The conduct took place in the context of and was associated with either an IAC or a NIAC (depending on the applicable provision)
  4. The perpetrator was aware of factual circumstances that established the existence of an armed conflict.

Notably, the crime does not require showing that the deprivation in question caused any particular form of civilian harm, suffering, or death. Rather, the focus is on the deprivation of indispensable objects and the intent to starve civilians.

In what follows, I outline whether there is reason to believe that the war crime could attach in the context of Tigray and identify some of the interpretive challenges a starvation crime prosecution would face. Needless to say, any individual case would need to be evaluated on its merits, following a comprehensive review of all relevant and admissible evidence.

Armed Conflict

It is clear that there is at least a non-international armed conflict in Tigray. There has been intense violence since November 2020 between an organized armed group (the TDF/TPLF) and the armed forces of two states (Ethiopia and Eritrea), among other actors. The TDF/TPLF is trained, well-armed, broad in its membership, and capable of controlling territory and engaging in coordinated military operations outside its area of control. The fighting has led to thousands of deaths and the displacement of 1.7 million, has drawn in both Ethiopian and Eritrean troops, and has elicited the scrutiny of the Security Council. On the relevance of these factors in establishing the existence of a NIAC, see, e.g., here (paras. 531-538). Although the legal question of when conflicts end (paras. 485-496) remains unsettled, it is clear that the conflict between the TDF/TPLF and the Ethiopian government has not yet ended. The TDF/TPLF is holding territory against the will of the federal government and continues to engage in belligerent detention; the government is an enduring posture of siege with respect to Tigray; the recent cease-fire was unilateral and temporary; and troops on both sides remain mobilized or engaged in fighting.

A war of self-determination against alien occupation, colonial domination, or a racist regime can qualify as an IAC underarticle 1(4) of Additional Protocol I, to which Ethiopia is party. However, it is unlikely that Ethiopia falls into any of those categories, and the TPLF has denied that it seeks full secession. The group does not appear to have made a unilateral declaration under article 96(3) of the Protocol (compare here) and its most recent demands include that the federal government “accept the fact that the constitution [of the Federal Democratic Republic of Ethiopia] is the sole basis for any negotiation going forward [and] enable the democratically elected Government of Tigray, with all its powers, and constitutional responsibilities, to resume its regular work.” An IAC classification under article 1(4) would anyway be specific to the rules of the Protocol, which does not classify starvation methods as a grave breach (art. 85). In the context of evaluating war crimes, then, the conflict may be presumed to be a NIAC.

The actions discussed below appear clearly to be “associated” with the conflict in that the conflict has played a substantial part in the decisions to engage in siege and other deprivation and shaped the purposes for which those acts are undertaken. They appear to have been undertaken by the military, against combatants and civilians of the opposing party (and possibly civilians of the same party), with a view to contributing to the goals of the military campaign (compare ICTY IT-96-23 paras. 58-60).


The central objective element of the crime is that the perpetrator must have deprived civilians of objects indispensable to their survival. Such objects clearly include “foodstuffs, agricultural areas for the production of foodstuffs, crops, livestock, drinking water installations and supplies and irrigation works,” as these are specified in the underlying rule of IHL (art. 14 of Additional Protocol II). However, this list is indicative, rather than exhaustive, so it could also include medical supplies and other essentials. There is no specification of the forms of deprivation that count, other than the clarification in the Rome Statute provisions that impeding relief supplies is included as a form of deprivation (Resolution ICC-ASP/18/Res.5 annex I). Preventing or disrupting coping strategies (pp. 10-12) or disrupting the viability of caring for (and thus feeding) children could also qualify. The primary forms of deprivation of objects indispensable to survival that would not qualify would be the deprivation of objects whose sustenance would go exclusively to combatants or deprivation unrelated to the denial of sustenance and not likely to threaten starvation or forced movement. That much, at least, is specified in the underlying IAC rule of IHL (art. 54 of Additional Protocol I).

Focusing exclusively on objects included in the indicative list above, there is significant evidence of the deprivation of objects indispensable to survival in the conflict in Tigray to date. In the earlier stages of the conflict, this included (pp. 24-52) the denial of access to certain areas, the looting or destruction of food, crops, livestock, and farm and medical equipment, the prevention of coping strategies, asset-stripping, attacks on humanitarian workers, and restrictions on movement that prevented planting and other essential activities. In the ongoing siege, it includes:

The standard view is that the denial of humanitarian access does not violate IHL if it is done for “valid reasons” or on “good grounds” and not “arbitrary or capricious” ones (here para. 2805; here para. 4885). However, denial that satisfies the mens rea of the war crime of starvation would almost certainly be arbitrary by definition (para. 51) due to the reasons for, or consequences of, the refusal of access. As such, arbitrariness need not be examined here. Separately, in light of ambiguity as to the authorship of the recent attack on the aid convoy in Afar, it is worth emphasizing that the criminal prohibitions discussed here apply to the deprivation of essential objects to any civilians as long as the elements of the crime are satisfied. This is true regardless of whether the affected civilians are affiliated with the force engaged in that deprivation, with its adversaries, or with neither.


The critical question then is whether those involved in the siege or other deprivation actions “intended to starve civilians as a method of warfare.” This is, of course, a question that requires individualized analysis. Nonetheless, several points are worth clarifying as investigators and others consider the possible criminality of the activities underlying the current crisis.

A Purpose Requirement?

On the most restrictive interpretation of what it means to “intend[] to starve civilians as a method of warfare,” this element is satisfied only when the perpetrator acts with the purpose of weaponizing the civilian harm or death associated with starvation with a view to furthering the war effort. Adopting this premise, some may argue that the purpose of the siege is to cut off and coerce the adversary TDF/TPLF forces, not to starve the civilians of Tigray – it is just that in a siege, the impact on civilians and combatants cannot be separated. A second line of argument, adopted publicly by the Ethiopian government, is that the obstruction of even humanitarian aid is designed to prevent the transfer of objects other than those indispensable to survival, including weapons and other contraband. There is reason to be skeptical of each of these lines of argument.

First, the siege ought to be evaluated in context. Prior to the eviction of federal and Eritrean forces in late June, there werereports of indiscriminate and comprehensive door-to-door looting, the burning of crops in door-to-door raids, the comprehensive and targeted destruction of water pumps, the forcing of owners to kill their own livestock, and attacks on humanitarian facilities and personnel. These appear, particularly in combination, to indicate action organized around the objective of starving the civilians of Tigray. That context may shed light on the purpose of ongoing impediments to humanitarian relief, certainly insofar as the same actors are involved. Notably, contextual factors are often important in establishing the purposive intent necessary for genocide (e.g. ICTY IT-98-33-T para. 580).

Second, and more fundamentally, denying the delivery of essentials to Tigray with a view to depriving the TDF/TPLF forces of sustenance would entail the purposive denial of sustenance to the population of Tigray. In other words, using the blunt tool of encirclement deprivation without control over what occurs in Tigray, the only way to starve out the TDF/TPLF would be to starve the population within which it is ensconced. Critically, however, the mere fact of TDF/TPLF presence in (or even control over) Tigray does not strip the overall population of its civilian character (e.g. art. 50(3) Additional Protocol I; ICTY IT-95-5/18-Tpara. 4610 n. 15510). As such, using encirclement deprivation to starve out the TDF/TPLF would entail purposively starving a civilian population as a method of warfare.

Similarly, the blanket prevention of objects entering Tigray cannot be defended as aimed at the impeding of weaponry and other contraband. When consignments are blocked without any meaningful effort to discriminate according to those features, the operation as a whole includes both the purposive impediment of military supplies and the purposive deprivation of objects indispensable to civilian survival. By way of analogy, the ICC Appeals Chamber has confirmed that attacks directed against both civilians and combatants without distinction count as attacks directed against civilians (paras. 418, 424, 491). Given that all currently available routes into Tigray go through Ethiopia or Eritrea, those in charge of those routes have the capacity to use more tailored means to prevent the delivery of weaponry without blocking objects indispensable to survival.

Third, it is important to note that permitting occasional humanitarian deliveries would not itself rebut a finding of purposive starvation, if the overall siege is clearly structured to deprive the encircled population of essential supplies. Notably in this respect, the limited deliveries into Tigray since June 28 fall woefully short of what is needed to avert mass starvation. By way of comparison, the International Criminal Tribunal for the former Yugoslavia found (para. 31) that the decision not to kill Bosnian Muslim women alongside the men and boys slaughtered at Srebrenica was compatible with holding the genocidal intent to destroy the Bosnian Muslims of Srebrenica (as a substantial part of the Bosnian Muslim group) because the former decision could be “explained by the Bosnian Serbs’ sensitivity to public opinion.” Committing to allow humanitarian delivery into Tigray can also assist with public relations. The key is how it features in the overall system of encirclement.

The Sufficiency of Oblique Intent

The discussion thus far has assumed the necessity of establishing purpose. However, a good case can be made that the war crime of starvation of civilians as a method of warfare does not require purposive intent at all. The ICC Statute (art. 30) defines “intent” with respect to a consequence to include either purposive intent or oblique intent. The latter entails the perpetrator acting with a virtual certainty that the consequence in question will occur, whether or not she seeks that result (here, paras. 447-450). Although that definition is technically applicable only to the default mens rea standards, crime-specific uses of “intentionally” should be read in light of that general meaning unless there is a good reason not to do so (ICC-01/05-01/13-2275-Red para. 677). Certainly the Rome Statute’s inclusion of oblique intent as “intent” is far from anomalous.

Looking to the underlying IHL framework, article 54 of Additional Protocol I is explicit in prohibiting the deprivation of indispensable objects when that may be expected to cause the starvation of civilians, even if the latter is not the purpose of the conduct (subpara. 3(b)). Moreover, as the ICRC Commentary to article 18 of Additional Protocol II notes (para. 4885), “If the survival of the population is threatened and a humanitarian organization fulfilling the required conditions of impartiality and non-discrimination is able to remedy this situation, relief actions must take place. … a refusal [without good grounds] would be equivalent to a violation of the rule prohibiting the use of starvation as a method of combat.”

As was emphasized at the Security Council earlier this month, the conditions in Tigray at this moment are such that even the partial denial of humanitarian relief portends starvation within the civilian population. With 400,000 people already suffering famine and 1.8 million on the brink thereof, it is difficult to imagine a context in which the denial of humanitarian relief would lead to starvation with a greater certainty than is applicable to Tigray today. Those who impede the delivery of indispensable objects with awareness of that context and without a clear justification are very likely to satisfy the threshold for oblique intent, whether or not it can be shown that their purpose is to weaponize the suffering of civilians.


Given the untested boundaries of the crime, the pursuit of a starvation-of-civilians case related to Tigray would raise challenging interpretive issues. However, this should not deter scrutiny of what is happening in the region. On the contrary, whichever interpretation gains traction, there is good reason to believe that extraordinary deprivation being inflicted on the people of Tigray could satisfy the criminal threshold for starvation of civilians as a method of warfare. Whether or not the more immediate violence of previous months returns, the siege of Tigray is producing precisely the kind of gradual and often-obscured violence that the starvation crime is designed to address. It is essential that region be opened to comprehensive humanitarian relief.

(Author’s note: Many thanks to Alex de Waal and Chidi Anselm Odinkalu for conversations and comments on an earlier draft.)

IMAGE: Members of the Amhara militia, that combat alongside federal and regional forces against northern region of Tigray, ride on the back of a pick up truck in the city of Gondar, on 08 November 2020. (Photo by EDUARDO SOTERAS/AFP via Getty Images)