Is There a Policy of Starvation in Gaza? Reading the Treaty Standard, the Determinations Issued and Each Party’s Own Counts

No court has issued a judgment on whether starvation has been used as a method of warfare in Gaza, and the determinations stating that it has were issued by five non-judicial bodies in dated publications between December 2023 and August 2025. Anyone following the war has seen the two flat statements sitting beside each other: that Gaza is being deliberately starved, and that no policy of starvation exists. Both are on the record, said by named people within weeks of each other. What the record separates, and what the two statements collapse, is the deprivation from the intent: the Integrated Food Security Phase Classification classified famine in Gaza in August 2025, having projected in July 2025 that at least 500,000 people were in Phase 5, its catastrophe level, while COGAT records 1,778,363 metric tons of food entering Gaza between 10 October 2025 and 7 June 2026 on its own count.

Key Findings

• The treaty text makes intent the load-bearing element: Rome Statute Article 8(2)(b)(xxv), adopted in 1998, defines the war crime as “intentionally using starvation of civilians as a method of warfare by depriving them of objects indispensable to their survival, including wilfully impeding relief supplies.”
• The Integrated Food Security Phase Classification classified famine in Gaza in August 2025, and its July 2025 projection placed at least 500,000 people in Phase 5, catastrophe. The classification grades food-security outcomes against thresholds and does not assess intent or assign responsibility.
• COGAT reports 1,778,363 metric tons of food entering Gaza between 10 October 2025 and 7 June 2026, its own count, with tonnages resting on self-declaration by suppliers rather than weighing at the crossings. Israel’s Ministry of Foreign Affairs counted over 70,000 tons of aid entering by 16 December 2023, also its own count.
• On 9 October 2023 Israel’s then defence minister announced a “complete siege” of Gaza; two months later the same government published aid-entry tonnages. Both are the party’s own record.
• The IDF Military Advocate General’s Corps reported 74 criminal investigations opened as at 3 August 2024, in five stated categories: detainee deaths, detainee mistreatment, pillaging, illegal use of force, and destruction of civilian property. No category concerning the restriction of humanitarian supplies appears among them.

Has any body determined that starvation in Gaza is policy?

Five bodies have published a determination that starvation in Gaza was used as a method of war, and none of them is a court. Human Rights Watch issued the first in December 2023, B’Tselem followed in April 2024, the UN Special Rapporteur on the right to food in July 2024, the UN Special Committee to Investigate Israeli Practices in November 2024, and Amnesty International in 2025.

The wording in each is explicit. The UN Special Committee stated in November 2024 that Israel was “using starvation as a method of war.” The Special Rapporteur on the right to food wrote in July 2024 that Israel had “made its intentions to starve everyone in Gaza explicit.” Amnesty International stated in 2025 that it had “compelling evidence” that the starvation of Palestinians in Gaza is a “deliberate policy.” B’Tselem published its April 2024 document under the conclusion that the war crime of starvation was being committed, and Human Rights Watch in December 2023 and again in April 2024 described starvation used as a method of warfare and imposed on the population.

Each of those is an organizational or mandate-holder determination: a body applying its own evidentiary standard, publishing its own conclusion, outside any proceeding in which the underlying facts were contested and tested. That is a different category of document from a judgment, and the distinction is not a matter of how strongly a conclusion is worded.

Against them stands the position of the government named. Prime Minister Benjamin Netanyahu stated in July 2025 that there is “no policy of starvation in Gaza” and “no starvation in Gaza,” while also stating that humanitarian aid is permitted to enter. Israeli bodies publish tonnage counts in support of the second half of that position, examined below.

Those two sets of statements are the whole of the dispute as it circulates. The rest of this report takes the pieces apart: the standard a policy finding has to meet, the statements offered as evidence of intent, what the famine monitors actually classified, what Israel’s own records report, where the measurement is contested, and what is absent from the record entirely.

Is it often said that Israel deliberately starves Gaza?

It is often said that Israel is deliberately starving Gaza, and six dated publications between December 2023 and August 2025 state that conclusion in those terms. Each was issued by a non-governmental organization or a UN mandate holder applying its own evidentiary standard, and each turns on intent, the element that separates a famine from a policy.

The six are Human Rights Watch in December 2023 and again in April 2024, B’Tselem in April 2024, the Special Rapporteur on the right to food in July 2024, the UN Special Committee in November 2024, and Amnesty International in 2025. The documents rest on overlapping material: witness and survivor testimony, records of what entered the territory and when, published statements by Israeli officials, and each body’s reading of the aggregate outcome.

The directional part of the claim is supported in the sense the claim is usually made: multiple named bodies, over two years, in published documents, reached the conclusion that the deprivation was intended, and none of them has withdrawn it. A reader who has encountered the claim as a consensus has encountered something real.

The reframe is the denominator. Six documents from five bodies is a count of assertions, not a change in the kind of evidence available. Adding a seventh organization stating the same conclusion does not convert an organizational determination into an adjudicated finding, because the evidentiary route is the same in each case: a body assembling material, applying its own threshold, and publishing. Over the same period, Israeli government bodies published counts and statements reaching the opposite conclusion, and Israeli research institutes published objections to the underlying measurement. Both sets are in the record; neither has been tested against the other in a proceeding.

One further distinction runs through the whole subject and is worth stating before the evidence sections. The claim contains two separable propositions: that Gaza experienced deprivation at famine levels, and that the deprivation was policy. The first is a measured outcome with a classification behind it. The second is a question about intent, and the two are answered by different bodies using different methods. The same separation governs what each body counts on civilian harm more broadly.

What must be shown before starvation counts as a policy?

The controlling text is Rome Statute Article 8(2)(b)(xxv), adopted in 1998, which defines the war crime as “intentionally using starvation of civilians as a method of warfare by depriving them of objects indispensable to their survival, including wilfully impeding relief supplies as provided for under the Geneva Conventions.” Intent sits inside the definition, not alongside it.

Two earlier instruments supply the underlying prohibitions. Geneva Convention IV of 1949, Article 23, obliges a party to allow the free passage of consignments of foodstuffs, clothing and tonics intended for children, expectant mothers and maternity cases, subject to stated conditions on diversion and control. Additional Protocol I of 1977 states at Article 54(1) that “starvation of civilians as a method of warfare is prohibited,” and at Article 54(2) prohibits attacking, destroying, removing or rendering useless objects indispensable to the survival of the civilian population, naming foodstuffs, agricultural areas, crops, livestock, drinking water installations and irrigation works. Article 70 governs relief actions and the passage of consignments.

The distance between a prohibition and a finding is a matter of thresholds, and the Rome Statute sets those out. Article 58 permits a pre-trial chamber to issue an arrest warrant where there are “reasonable grounds to believe” a person committed a crime within jurisdiction. Article 66 sets the standard for conviction: the accused is presumed innocent, and the court must be convinced of guilt beyond reasonable doubt. The International Criminal Court issued arrest warrants in November 2024 in the Gaza matter, a decision taken at the Article 58 threshold. That threshold is the lowest at which the institution acts on a charge, and it is separated from a determination of guilt by the whole of a trial.

The practical consequence is that an outcome measurement, however severe, does not by itself satisfy the definition. Article 8(2)(b)(xxv) requires the deprivation to be intentional, which is why the famine classification and the policy question are answered in different documents by different bodies.

Which statements are cited as evidence of intent?

One statement is cited more than any other. On 9 October 2023 Israel’s then defence minister announced a “complete siege” of Gaza, stating “no electricity, no food, no fuel, everything is closed.” The Special Rapporteur on the right to food in July 2024, B’Tselem in April 2024 and Amnesty International in 2025 each treat that announcement as the moment the intention was stated openly.

The statement is a primary record of what a minister of the government said, on a dated occasion, and it is the government’s own. That is what gives it weight in the documents that rely on it, and it is also the limit of what it establishes on its own: it records an announced order, not the operative directives that followed it, and it does not record what quantity of what entered the territory on any subsequent date.

The same government’s published counts are the answer it offers. Israel’s Ministry of Foreign Affairs reported that over 70,000 tons of aid had entered Gaza as at 16 December 2023, including over 6,500 tons of medical supplies, alongside over 750 patients who exited for care overseas and approximately 9,500 foreign nationals and dual citizens evacuated with Israeli assistance. Those are the ministry’s own counts, published by a party to the conflict, and each is written as a floor or an approximation.

So the public record holds a stated total siege on 9 October 2023 and, from the same government, reported tonnage two months later. What it does not hold, in either direction, is the set of documents that would connect them: the internal instructions issued to the bodies controlling the crossings, and the distribution records showing what reached whom. The determinations that treat the October statement as proof of intent are reading an announced order forward; the counts that answer them are reading a later month backward. Neither is the intervening record.

Israel also publishes a general position on precautions. The IDF states that it warns civilians by telephone calls, text messages and leaflets, publishes numbered evacuation-area maps, and aborts strikes when civilians are present, and it publishes no counts for any of those measures. The same asymmetry applies to the aid question: a stated practice with no published magnitude cannot be weighed against a measured outcome.

What did the famine monitors classify, and when?

The Integrated Food Security Phase Classification classified famine in Gaza in August 2025, and its July 2025 projection placed at least 500,000 people in Phase 5, the catastrophe level its own five-phase scale defines by starvation, destitution and death. The World Health Organization stated in August 2025 that famine had been confirmed in Gaza for the first time.

The classification is the closest thing in this subject to a settled measurement, and its method is what makes it citable. The system is used by governments, UN agencies and relief organizations to grade food insecurity against published thresholds covering household food consumption, acute malnutrition in children, and mortality. It is a consensus classification of observed conditions, run on survey and screening inputs collected in the territory. Peer-reviewed medical literature published in 2025 documents the same picture from the clinical side, recording hunger and malnutrition at levels the authors describe as unprecedented for the territory.

The limit is as important as the finding, and the classification states it itself. The system assesses food-security outcomes. It does not assess intent, it does not attribute causation to a party, and it issues no legal characterisation. A famine classification and a determination of policy are answers to different questions, and the classification cannot supply the element that Article 8(2)(b)(xxv) requires.

That is why the classification appears on both sides of the public dispute without contradiction. It is cited by the bodies that determined a deliberate policy, as the outcome half of their conclusion; and it is compatible, as a measurement, with any account of how the outcome arose. A reader who takes the classification as the answer to the policy question has taken an outcome measurement for an intent finding, which is the single most common move in this subject and the one the classification’s own scope note forecloses.

One further scope point: the classification covers the population present in the Gaza Strip at the time of assessment, and the size of that population is itself a moving figure. The Palestinian Central Bureau of Statistics reported about 2.13 million people in Gaza at the end of 2025, a decline of approximately 254,000 which it states is 10.6% against its pre-aggression estimates. It separately records roughly 2.2 million living in Gaza on the eve of the war, which is a different quantity: the 254,000 and the 10.6% fix their own base at about 2.384 million, so the percentage is not the distance between 2.2 million and 2.13 million.

What does Israel’s own record state about aid entering Gaza?

COGAT records 1,778,363 metric tons of food entering the Gaza Strip between 10 October 2025 and 7 June 2026, its own operational count. COGAT is a unit of the Israeli Ministry of Defense and a party to the events it describes, and its report states that tonnages rest on self-declaration by humanitarian organizations and private-sector providers rather than on weighing at the crossings.

Source: each body’s own published count, as reported by that body. None is independently verified. The periods and the categories differ (food only in the first row, all aid in the other two), so the three figures are not one series and must not be compared or summed.

COGAT reports further figures on the same basis: more than 70,000 cubic meters of water per day facilitated through external water lines and desalination facilities, excluding local wells, and 18,000 tons of medical supplies and medicines since the ceasefire began on 10 October 2025. The Israel Defense Forces reported 7,800 tons cumulatively delivered by the aerial route as at 19 August 2025, and 23,000 tons overseen in the week ending that date.

Two limits travel with every figure in this section. First, each is a count of entry, not of distribution, and none of these publications reports what reached which households. Second, the figures written as “over” or “approximately” are floors and approximations in their own sources and may not be restated as exact totals.

The position built on those counts is the one stated at the top of the government. Netanyahu stated in July 2025 that there is no policy of starvation in Gaza. The counts above are the evidence the government publishes for the aid half of that statement; they are its own reporting, and they do not address the intent element separately.

Which analyses reject the famine and policy findings?

Two Israeli research institutes have published the objection in dated documents. The Institute for National Security Studies published “The Misleading Reports of the UN Over Famine in Gaza” in 2025, arguing that the data inputs and the application of the classification thresholds do not support the finding as issued. The Jerusalem Center for Security and Foreign Affairs published its own 2025 essay contesting the atrocity framings applied to the war.

The disagreements are separable, and running them together is what makes the public dispute unreadable. Three distinct objections circulate. The first concerns inputs: whether household surveys, child malnutrition screening and mortality estimates collected during active hostilities and mass displacement can support the precision the classification’s thresholds require. The second concerns application: whether the thresholds, once the inputs are accepted, were met on the evidence presented. The third concerns inference: whether an outcome, however measured, establishes the intent that a policy finding requires.

The Institute for National Security Studies argues the first two. The Jerusalem Center’s 2025 essay operates mainly on the third, and its own arithmetic illustrates the limits of the exercise: it derives a ceiling of at most 41,000 civilians in Gaza who have died of war-related causes by subtracting estimated natural deaths and Israel’s count of Hamas fighters killed from the Gaza Ministry of Health total, and the page itself states that neither of those inputs has been independently confirmed. That is the publisher’s own calculation on figures supplied by parties to the conflict, stated as a ceiling rather than a count.

Analysis running the other way is also on the record. The Guardian published a reconstruction in July 2025 setting recorded aid entry against the caloric requirement of the population, reaching the conclusion that the deprivation followed from the quantities allowed in. That is a named press analysis of published tonnage and requirement figures, not a peer-reviewed measurement, and it is cited here as the outlet’s own reconstruction.

What none of the three objections reaches, and what none of the analyses on either side substitutes for, is an adjudication on contested facts. The dispute about the measurement and the dispute about the inference are both live, and they are argued in publications rather than resolved in a forum with the power to test them.

What would settle the policy question, and does it exist?

No court judgment on starvation as a method of warfare in Gaza appears in the public record. The International Criminal Court matter stood at the arrest-warrant stage in November 2024, a decision taken at the Rome Statute’s Article 58 “reasonable grounds to believe” threshold, well below the Article 66 requirement of proof beyond reasonable doubt. International Court of Justice proceedings concern provisional measures and the genocide question rather than a starvation-policy ruling.

Israel’s own accountability machinery publishes counts, and the categories are as informative as the totals. The IDF Military Advocate General’s Corps reported 74 criminal investigations opened as at 3 August 2024: 44 concerning the death of detainees held by the IDF, 8 alleged detainee mistreatment, 13 alleged pillaging or stealing, 6 alleged illegal use of force, and 3 alleged destruction of civilian property without military necessity. Over 1,000 incidents had been referred to the General Staff’s Fact-Finding and Assessment Mechanism for factual assessment. No category concerning the restriction of humanitarian supplies appears among the five published.

The Corps also publishes the standard that produces those counts, in its own words: in armed conflict, “death or injury to a civilian or harm to civilian objects, while tragic and regrettable, in most cases do not by themselves indicate a reasonable suspicion of criminal conduct, as they can be explained as reasonable aftermath of lawful operational activity.” The investigation count is a function of that threshold. All of these are the IDF’s own counts of its own processes, and they record investigations opened, not findings, indictments or convictions.

What a determination on the policy question would require is documentary: the internal directives governing what was permitted through the crossings and when, the distribution records showing what was delivered and to whom, and a forum in which those materials are tested against contested evidence. None of the three is in the public record. That absence is the state of the record, not a stage of it: the classification of the outcome exists and is dated, and the corresponding adjudication of intent does not. The same structural gap governs the genocide standard, where the convention text sets an intent element the outcome counts cannot supply.

Methodology and limitations

This report reads four kinds of material: treaty text, dated determinations published by non-judicial bodies, food-security classifications, and each party’s own published counts. It does not measure conditions in Gaza.

Treaty texts are cited from the instruments themselves: the Rome Statute of the International Criminal Court (1998), Articles 8, 58 and 66; Geneva Convention IV (1949), Article 23; and Additional Protocol I (1977), Articles 54 and 70. These are primary records and are quoted rather than paraphrased where the wording is load-bearing.

Determinations by Human Rights Watch (December 2023, April 2024), B’Tselem (April 2024), the UN Special Rapporteur on the right to food (July 2024), the UN Special Committee to Investigate Israeli Practices (November 2024) and Amnesty International (2025) are named and dated as the conclusions of those bodies, reached on their own evidentiary standards. They are not treated as adjudicated findings, and none is used as this report’s measurement of any quantity.

The Integrated Food Security Phase Classification is a multi-partner classification system grading food-security outcomes against published thresholds. Its August 2025 famine classification and its July 2025 Phase 5 projection are cited as classifications of outcomes. The system does not assess intent or attribute responsibility, and the figures cited here carry that scope. The peer-reviewed 2025 literature on famine conditions in Gaza (PubMed Central, PMC12118885) is cited as clinical documentation of deprivation levels, not of causation.

Israeli figures come from COGAT (2026), the Israel Defense Forces (2025), the Israel Ministry of Foreign Affairs (2023) and the IDF Military Advocate General’s Corps (2024). Every one is a party’s own count of its own activity, published by a party to the events described, and none is independently verified. COGAT states that its tonnages rest on self-declaration by suppliers. Counts written as “over” or “approximately” are floors and approximations. The three aid totals cover different periods and different categories and are not a single series.

Where the literature disagrees, both sides are named: the Institute for National Security Studies (2025) contests the classification’s inputs and threshold application; the Guardian (2025) published a reconstruction of aid tonnage against caloric requirement reaching the opposite conclusion; the Jerusalem Center for Security and Foreign Affairs (2025) argues on the inference from outcome to characterisation, using its own arithmetic on figures its own page states are unconfirmed.

Public statements by Israeli officials, including the 9 October 2023 announcement of a complete siege and the July 2025 statement that no policy of starvation exists, are cited as party statements on the record, dated, and not as findings.

One absence is recorded rather than filled: no judicial determination on whether starvation has been used as a method of warfare in Gaza appears in the public record, and the internal directives and distribution records that a determination would test are not public.

Conclusion

The two statements a reader has seen side by side are both on the record, and they answer different questions. The deprivation has been classified: the Integrated Food Security Phase Classification graded famine in Gaza in August 2025, with at least 500,000 people in its catastrophe phase on the July 2025 projection. The intent has not been adjudicated: the determinations that call it policy were issued by five bodies applying their own standards, and no court has ruled.

The rest of the record is a set of documents that do not meet. The treaty text puts intent inside the definition of the crime, so an outcome measurement, however severe, cannot complete it. A minister announced a complete siege on 9 October 2023, and the same government published tonnage counts two months later; both are that party’s own record, and the directives between them are not public. COGAT records 1,778,363 metric tons of food entering across an eight-month ceasefire window, self-declared and unverified, while publishing nothing on distribution. Israel’s own military prosecution reports 74 criminal investigations opened in five categories, none of them concerning humanitarian supply, under a threshold it states plainly. Every one of those is a real, checkable fact, and none of them is the fact the question turns on.

That leaves a gap with a shape. The severity is measured and dated; the intent is asserted, denied, and untested. Into a gap like that, two settled answers move in, each carrying the confidence of a finding neither has. And the harder question sits at the far end of that traffic: when the conduct of a war is described as a deliberate policy of starvation, and the description circulates as adjudicated when it is not, what happens to the language used about Jews as such, and does the unresolved space itself become one of the inputs? The two dated sentences that opened this, weeks apart, from a government and from an organization, are still the whole of what the public record settles.

Frequently Asked Questions

Does a famine classification identify who is responsible?

No. The Integrated Food Security Phase Classification grades food-security outcomes against published thresholds on household food consumption, acute malnutrition and mortality, and its August 2025 classification of famine in Gaza is a statement about conditions. Attribution of cause and any assessment of intent sit outside its method.

Where do the circulating malnutrition death figures come from?

The Gaza Ministry of Health reported in April 2024 that 32 people, including 28 children, had died of malnutrition and dehydration at hospitals in northern Gaza. That is the ministry’s own count, published by a party to the conflict, and outlets that carry the figure are relaying it rather than measuring it.

Why do Israel’s published aid totals not form a single series?

Because each covers a different window and a different category. COGAT’s 1,778,363 metric tons is food only, from 10 October 2025 to 7 June 2026; the IDF’s figure of over 1.9 million tons as at 19 August 2025 is all aid since the start of the war; the Ministry of Foreign Affairs counted over 70,000 tons of all aid to 16 December 2023.

How does COGAT compare its recorded food entry with what Gaza needs?

COGAT states its recorded food entry over the ceasefire period was nearly threefold the World Food Programme’s monthly food requirement benchmark. Only the numerator in that ratio is COGAT’s own count; the benchmark is the WFP’s, and COGAT does not report distribution against it.

What does international law require of a party controlling access to food?

Geneva Convention IV of 1949, Article 23, obliges a party to allow free passage of food consignments intended for children, expectant mothers and maternity cases, subject to stated conditions on control and diversion. Additional Protocol I of 1977, Article 54(2), prohibits destroying or rendering useless foodstuffs, crops, livestock and drinking water installations.

What do Israel’s records report on water and medical supplies?

COGAT reports more than 70,000 cubic meters of water per day facilitated through external water lines and desalination facilities, excluding local wells, and 18,000 tons of medical supplies and medicines since 10 October 2025. Both are COGAT’s own counts and are not independently verified.

How many people were living in Gaza at the end of 2025?

The Palestinian Central Bureau of Statistics reported about 2.13 million people in Gaza at the end of 2025, a decline of approximately 254,000 against its pre-aggression population estimates, which it states is a decrease of 10.6%. It reports roughly 2.2 million living in Gaza on the eve of the war, and attributes the change to war-related deaths, missing persons, reduced births and forced out-migration.

Sources

• Catastrophic famine in Gaza: unprecedented levels of hunger, 2025. PubMed Central, PMC12118885. Peer-reviewed.
• COGAT (Coordinator of Government Activities in the Territories), 2026. Humanitarian Situation Report During the Ceasefire. gaza-aid-data.gov.il. Official record, an Israeli Ministry of Defense unit’s own reported counts.
• Gaza Ministry of Health, 2024. Reported malnutrition and dehydration deaths at hospitals in northern Gaza, April 2024. Party’s own count.
• Geneva Convention IV, 1949, Article 23; Protocol Additional to the Geneva Conventions (Additional Protocol I), 1977, Articles 54 and 70. Treaty text. Primary record.
• Institute for National Security Studies, 2025. The Misleading Reports of the UN Over Famine in Gaza. inss.org.il. Research institute, not peer-reviewed.
• Integrated Food Security Phase Classification, 2025. Gaza Strip famine classification (August 2025) and Phase 5 projection (July 2025). ipcinfo.org. Multi-partner food-security classification system.
• Israel Defense Forces, 2025. Clearing Up the Misconceptions: Everything You Need to Know About the Humanitarian Aid Situation in Gaza. idf.il. Official record, a party’s own reported counts.
• Israel Defense Forces, Military Advocate General’s Corps, 2024. Addressing Alleged Misconduct in the Context of the War in Gaza, updated 3 August 2024. idf.il. Official record, a party’s own reported counts.
• Israel Ministry of Defense, 2023. Statement of 9 October 2023 announcing a complete siege of the Gaza Strip. Party statement on the record.
• Israel Ministry of Foreign Affairs, 2023. Israel-Hamas Conflict 2023: Humanitarian Efforts, as at 16 December 2023. gov.il. Official record, a party’s own reported counts.
• Jerusalem Center for Security and Foreign Affairs, 2025. The Death of Genocide. jcfa.org. Research institute, not peer-reviewed.
• Netanyahu, Benjamin, 2025. Public statements of July 2025 that there is no policy of starvation and no starvation in Gaza. Party statement on the record.
• Palestinian Central Bureau of Statistics, 2025. A Brief on the Status of the Palestinian People at the End of 2025. pcbs.gov.ps. Official record, national statistical office.
• Rome Statute of the International Criminal Court, 1998, Articles 8, 58 and 66. Treaty text. Primary record.
• The Guardian, 2025. The mathematics of starvation: how Israel caused a famine in Gaza, 31 July 2025. theguardian.com. Named press analysis.
• World Health Organization, 2025. Famine confirmed for first time in Gaza, 22 August 2025. who.int. Intergovernmental agency statement.