COGAT recorded 1,778,363 metric tons of food entering the Gaza Strip between 10 October 2025 and 7 June 2026, a count the Israeli Ministry of Defense unit that coordinates the crossings built from tonnages self-declared to it rather than weighed, and one that is not independently verified. At the far end of the same delivery chain, the Site Management Cluster counted 136,000 households in Gaza’s displacement sites, 38% of all households it enumerated, that had received no shelter assistance in the six months before its data extraction of 16 May 2026. It is often said that Israel prevents Palestinians in Gaza from accessing humanitarian aid. The United States Congress put a test for that conduct on the statute book in 1996, and the executive assessment applying it was transmitted to Congress on 10 May 2024 and cannot be read on any United States government host.
Key Findings
- United States law already names the conduct. The Foreign Assistance Act states, at section 620I, that no assistance shall be furnished under that Act or the Arms Export Control Act to any country when it is made known to the President that its government “prohibits or otherwise restricts, directly or indirectly, the transport or delivery of United States humanitarian assistance.”
- The bar carries an unconditioned waiver and names no adjudicator. Subsection (b) lifts it wherever the President determines that furnishing assistance “is in the national security interest of the United States,” and subsection (c) asks only for notice of the intention, the effective date and the reasons.
- The assessment that applies the standard is not published. A White House memorandum of 8 February 2024 states that a written report shall go to six congressional committees within 90 days and once every fiscal year after that; the first was transmitted on 10 May 2024, and a search on 19 August 2026 found no copy on state.gov or on the archived 2021-2025.state.gov.
- 136,000 households in displacement sites, about 680,000 people and 38% of all households counted, had received no shelter assistance in the six months before the survey. The Site Management Cluster counted them across 1,591 active sites, on data extracted 16 May 2026.
- 7,700 households, about 38,500 individuals, were sleeping outdoors with no shelter at all, a condition the Site Management Cluster records in 19% of the sites it surveyed; about 1.5 million residents were living in roughly 246,000 tents, about six people to a tent designed for five.
What does United States law count as restricting humanitarian access?
Section 620I of the Foreign Assistance Act bars United States assistance to any country when it is made known to the President that the government of that country “prohibits or otherwise restricts, directly or indirectly, the transport or delivery of United States humanitarian assistance.” The United States Congress put that sentence on the statute book on 30 September 1996.
The Foreign Assistance Act states the rule in one subsection: “No assistance shall be furnished under this chapter or the Arms Export Control Act to any country when it is made known to the President that the government of such country prohibits or otherwise restricts, directly or indirectly, the transport or delivery of United States humanitarian assistance.” The Law Revision Counsel publishes it at 22 U.S.C. 2378-1, added by Public Law 104-208 and printed at 110 Stat. 3009-121, where the section runs to three subsections.
Three phrases carry the weight. “Prohibits or otherwise restricts” reaches conduct short of a formal ban. “Directly or indirectly” reaches conduct short of an order to whoever carries the goods. And “United States humanitarian assistance” fixes the object, so a restriction on a consignment funded by another government falls outside the words Congress used.
What the section leaves out is equally specific. It says nothing about international humanitarian law, nothing about civilian harm and nothing about proportionality. Its subject is the movement of goods, and its test is whether a government prohibited or restricted that movement.
That is the enacted definition of the conduct behind this report’s title, and the United States Code records it as reaching both the assistance channel and the arms channel, decades before the current war.
Who decides whether that standard has been crossed?
Section 620I names one decider and supplies no threshold. Its trigger is the phrase “when it is made known to the President,” and subsection (b) lifts the bar wherever the President determines that furnishing assistance “is in the national security interest of the United States.” Subsection (c) asks only for notice of the intention, the effective date and the reasons.
That notice runs to named congressional committees. The Foreign Assistance Act still names the Committee on International Relations, and the United States Code records that the House renamed it the Committee on Foreign Affairs by House Resolution 6 of 5 January 2007, carrying the change as an editorial note rather than a revision of the enacted words.
Read together, the three subsections describe a closed circuit. The section does not say what makes a restriction known. It names no agency to find facts, no court to review a finding and no standard of proof, and defines no term. The bar operates when the President treats the condition as met, the waiver when the President determines the national security interest requires it, and the committees receive notice afterwards.
The section therefore generates a visible artefact only when a determination is made or a waiver notice filed, and this report located neither, for any country, among these sources.
An absence of that kind is a fact about what has been published and nothing more. It is not evidence that the standard was met, and not evidence that a determination was avoided. The statute is drafted so that silence is its ordinary state, which is why the instrument alone cannot answer the question people ask of it.
What written assurance did the United States require in 2024?
A White House memorandum of 8 February 2024 states that the Secretary of State shall obtain written assurances that a recipient of United States defence articles will “facilitate and not arbitrarily deny, restrict, or otherwise impede, directly or indirectly, the transport or delivery of United States humanitarian assistance.” The memorandum places that requirement “in furtherance of supporting section 620I of the Foreign Assistance Act of 1961.”
The instrument is National Security Memorandum 20, and the Government Publishing Office records its official text in the Daily Compilation of Presidential Documents at DCPD-202400098. Section 1(a)(i) requires a second assurance, that transferred articles will be used “in accordance with international humanitarian law and, as applicable, other international law.”
The wording of the access assurance is close to the statute’s without being identical. Where the Foreign Assistance Act states “prohibits or otherwise restricts,” the memorandum states “arbitrarily deny, restrict, or otherwise impede,” and adds a positive duty to facilitate, covering both United States humanitarian assistance and government-supported international efforts to provide it.
What kind of instrument this is matters more than its wording. A national security memorandum is a presidential directive, not a statute. It imposes no obligation on any recipient country, creates no cause of action, and can be rescinded by a later President without any act of Congress.
The assurances are the recipient government’s own written representations about its own future conduct, so a report resting on them reports what a government undertook to do, not a measurement of what it did. The assurance is the input; the assessment described next is where the executive branch was to say what happened.
Where is the assessment that applies the standard?
Section 2(a) of the same memorandum states that a written report shall go to six named congressional committees “Not later than 90 days after the date of this memorandum, and once every fiscal year thereafter.” The Department of State transmitted the first such report on 10 May 2024. A search conducted on 19 August 2026 found no copy of it on state.gov or on the archived 2021-2025.state.gov.
Section 2(a)(ii) states that it shall carry “an assessment of any credible reports or allegations that such defense articles… have been used in a manner not consistent with international law,” and shall include “any determinations, if they can reasonably be made,” about whether the recipient country “has pursued appropriate accountability.” That is the executive branch applying its own standard to its own transfers, on a fixed schedule.
The memorandum does not require publication. Its duty runs to the six committees, so the absence of a public copy is consistent with the instrument rather than a departure from it. The unclassified text in circulation was posted by a third-party law-and-policy publication, not by the body that wrote it, and no later annual edition was located.
For that reason this report describes the May 2024 report and does not quote it. A document not read from its issuing body’s own host is described, dated and left unquoted, because quoting a text through a relayer is how a wording drifts.
The gap has a particular shape. The standard is enacted and public, the assurance requirement is quotable to its clause numbers, and the one document applying both to a named recipient reaches the public through somebody else’s posting, or not at all.
Israel recorded 1,778,363 metric tons of food entering during the ceasefire
COGAT recorded 1,778,363 metric tons of food entering the Gaza Strip between 10 October 2025 and 7 June 2026, alongside 18,000 tons of medical supplies since the ceasefire began. Both are its own reported counts, published by a party to the events and not independently verified.
The method travels with the figure, because the report carrying it states the method. COGAT publishes tonnages that rest on self-declaration by the humanitarian organisations and private-sector providers moving the goods, rather than on weighing at the crossings, so the total is a compilation of what carriers declared rather than a measurement taken by the recording body. That is a property of the count, not a complaint about it: no body publishes a weighed series for the same window.
What the figure measures is entry. It counts material that crossed into the Strip over a defined period, at the point of crossing, and it does so to the ton. It does not count distribution, and COGAT publishes no distribution series. It does not count what any household received, and it does not count losses between the crossing and the recipient, whether to damage, diversion, spoilage or the roads.
That distinction is the whole of the measurement problem here. A figure of this size answers a question about entry with precision and answers nothing about receipt. Reading it as an answer about access asks a crossing-point count to speak for a distribution system no single body measures end to end.
COGAT does set its own entry figure against a requirement benchmark published by another organisation, which is the subject of the next section.
Does Israel’s own report claim the entries met the need?
COGAT’s report makes one sufficiency claim: that recorded food entry over the same window ran at “nearly threefold” the World Food Programme’s monthly food requirement benchmark, alongside an approximately 72% decline in the Gaza food consumer price index between September 2025 and May 2026. It is published by a party to the events, and only the numerator of that ratio is its own.
The denominator belongs to another organisation. A monthly food requirement benchmark estimates how much food a population needs, on the estimating agency’s own method, so the ratio pairs one body’s count of entry with another’s estimate of need and moves with either. COGAT states the ratio without publishing the arithmetic behind it.
The price series carries a different limitation. A 72% decline in a food consumer price index is a market signal, and COGAT’s report identifies no source for the price data, so the collection method, the basket of goods and the geographic coverage are all unstated. Prices fall where supply reaches functioning markets, which makes such a series informative about market conditions and silent about households outside them.
This is the only sufficiency claim in the record assembled here. No other body among these sources assesses whether the volumes entering Gaza met the population’s requirement over the ceasefire window, in either direction.
That is worth stating plainly, because “is enough getting in” is the question most often asked about Gaza and the one the record answers least. A party to the events answers it, in its own terms, against a benchmark it did not set, and nobody else answers it at all.
What did Israel record entering before the ceasefire?
The Israeli Foreign Ministry reported more than 106,000 aid trucks carrying over two million tons of food, water, medicine and equipment entering Gaza since 7 October 2023, in a statement current to 14 September 2025, and more than 12,000 trucks since May 2025, 80% of them carrying food. Every figure is written as a floor, and every one is as reported by Israel.
The series has earlier points. The Israeli Ministry of Foreign Affairs recorded over 70,000 tons of aid and over 6,500 tons of medical supplies entering as of 16 December 2023. The Israel Defense Forces states, on a page date-stamped 19 August 2025, that over 1.9 million tons had been delivered since the start of the war, a figure written as a floor and one that cannot be restated as an exact 1,900,000.
These readings are not a time series and cannot be treated as one. They are separate statements by three arms of a single party, on different cut-off dates, in different units, some counting trucks and some counting tons, none defining categories the same way. Placing them on one axis would produce a trend nobody published and no method supports.
What they establish is narrower and still substantial. The party controlling the crossings has published its own dated running counts of entry throughout the war, and has pointed readers to its own aid-data site for detail. Those are the fullest entry-side record for the period before the ceasefire, and they share the property of the ceasefire-window figure: trucks and tons are units of entry, and a count of them is a count of what crossed a line.
How many people were counted living in displacement sites?
The Site Management Cluster counted 1,591 active displacement sites across the Gaza Strip, holding an estimated 1,708,822 individuals, on data extracted from its platform on 16 May 2026. Its enumerators visited every site on the cluster’s master list between 3 February and 19 April 2026, in a baseline verification rather than a point-in-time snapshot.
Fieldwork ran through 165 trained enumerators from twelve partner organisations across 116 neighbourhoods, targeting 1,856 sites and confirming 1,591 as active. The Site Management Cluster puts the population of those sites at 354,550 households. Makeshift sites, meaning informal, open-air and tented, are 83% of the total; about 11% are collective centres in schools and public buildings; the rest are scattered sites of fewer than ten households.
Two properties of the count travel with it. The Site Management Cluster states that the 1,591 “should be understood as a social count: A measure of how many organised displaced communities were identified and reached rather than a census of discrete physical locations.” And household size is imputed at five wherever disaggregated data was unavailable, which makes the household total derived rather than enumerated. Sex- and age-disaggregated data exist for only 253 of the sites.
The distribution is uneven, and the assessment publishes it by governorate: Khan Younis 834,547 people across 572 sites, Deir Al-Balah 424,125 across 430, Gaza 300,638 across 427, North Gaza 130,334 across 141 and Rafah 19,178 across 21. Those five figures sum to the reported total, which makes the breakdown a partition of the counted population rather than five independent estimates. A separate report covers the counts each body publishes for Gaza.
38% of households in displacement sites received no shelter assistance
In Gaza’s displacement sites, 136,000 households, about 680,000 people and 38% of all households counted, had received no shelter assistance in the six months before the Site Management Cluster’s field survey. That is a count of assistance received, taken at the site, at the opposite end of the delivery chain from a tonnage recorded at a crossing.
The Site Management Cluster counted 7,700 households, about 38,500 individuals, sleeping outdoors with no shelter at all, a condition present in 19% of the sites it surveyed. About 1.5 million residents were living in roughly 246,000 tents or tent-like shelters, an average of about six people to a tent against a standard emergency tent designed for five, and about 59,000 shelters housed more than eight people.
The cluster publishes all of this as estimates and states the limits itself. Household size is imputed at five where disaggregated data was unavailable. A single interview per site means the figures for persons with disabilities and for female-headed households are underestimates. The earliest data were about four months old at publication, and the Site Management Cluster states that its correlations are not causal.
It is worth being exact about what this measurement establishes. It records what households in enumerated sites reported having received across a six-month window. It identifies no cause and names no party, and it does not distinguish between assistance never funded, never sent, never cleared or never distributed.
Set beside a crossing-point tonnage, it establishes one thing firmly: the two ends of the delivery chain were measured by two bodies, using two methods, on two universes. Neither count is evidence about the other.
What is measured about water reaching households?
The WASH Cluster measured water insecurity in 82% of 342 assessed households, in a phone survey conducted between 2 and 15 March 2026 whose authors state that its findings “cannot be considered representative of the population in Gaza.” The share is widely relayed as a share of families in Gaza, and that relay overstates its base.
UNICEF relayed it in a mid-year 2026 humanitarian situation update, in the words “An estimated 82 per cent of families are water insecure,” which states a population share where the exercise assessed 342 households reachable by telephone.
The correction is the finding here, rather than the share itself. The number is real and was properly measured on the base it had. That base is small and selected by reachability, and its authors state that the results cannot be read as representative of Gaza’s population. A share of 342 assessed households and a share of Gaza’s families are two different quantities, and only the first was measured.
On the supply side, COGAT states that more than 70,000 cubic metres of water per day are facilitated into the Strip through external water lines and desalination facilities, excluding local wells, as at its July 2026 report. That is a party’s own count, published by a party to the events, and it measures water made available at the point of entry.
The two figures are not two readings of one quantity, and the difference runs through this whole report. One counts cubic metres entering a territory; the other classifies households by whether they can reliably obtain water. Between them sits a distribution network whose measured condition is the subject of the next section.
What condition are the systems that aid moves through in?
The Gaza Strip Rapid Damage and Needs Assessment reported more than 371,888 housing units destroyed or damaged and more than 50% of hospitals non-functional. The World Bank, the European Union and the United Nations published it jointly, covering October 2023 to October 2025 and combining remote sensing with ground-verified data validated by all three publishers.
The assessment puts three separate financial quantities and combines none of them. Physical infrastructure damages stand at US$35.2 billion and economic and social losses at US$22.7 billion, both backward-looking. Recovery and reconstruction needs over the next decade stand at US$71.4 billion, of which US$26.3 billion is required in the first eighteen months. The World Bank assessment puts the contraction of the economy at 84%.
The needs figure and the damage figure are not two readings of one loss. Damages price what has already been destroyed; needs price what rebuilding would cost over the decade ahead. Neither figure records money spent or money pledged, and the assessment never sums the two, because they answer different questions about different periods.
Why the condition of the system belongs in a report about access is a matter of plumbing. Aid arrives by road, is stored in warehouses, moves through clinics, bakeries, water points and schools, and ends in housing. When the assessment records more than 371,888 housing units damaged or destroyed and more than half of hospitals non-functional, the network that would carry a ton of flour from a crossing to a family is itself among the objects being measured.
What does peer-reviewed research measure about household access?
Hamamra and colleagues (2025) conducted 30 semi-structured interviews with displaced adult caregivers in the Rafah displacement camps, evenly split between women and men, reporting thematic saturation at the 25th interview. It is the only peer-reviewed study among these sources that asks households directly about access, and it carries no prevalence measurement of its own.
The study is qualitative by design. It codes interview transcripts into six themes, among them widespread malnutrition, disease in shelters, inaccessibility of healthcare and daily struggles for basic needs. It examines no patients, samples no population and reports no rate, so it is evidence about what a set of caregivers described, not about how common any of it was.
Every figure inside it belongs to another body, and the paper says so. UNICEF measured the under-two acute-malnutrition series it carries, and the Palestinian Central Bureau of Statistics counted 28 child deaths from malnutrition and dehydration in the first six months of the war. Neither is the authors’ own measurement, and neither should be re-attributed to the paper relaying it.
The authors state a causal conclusion of their own, that the crisis is “a deliberate outcome of the Israeli policies designed to undermine civilian survival,” and their title places the events during a genocide in Gaza. Both are the authors’ characterisations, drawn from interview themes rather than from a measured attribution of cause.
The finding this section records is an absence. No peer-reviewed study among these sources measures the prevalence of aid receipt among Gaza households, which is why the receipt half of this question rests on a humanitarian coordination body’s field enumeration and not on the academic literature.
Methodology and limitations
This report assembles four kinds of source and states what each measures.
The first is enacted United States law and a presidential directive, quoted from the United States Code and from the Daily Compilation of Presidential Documents. Those are instruments: they establish what a standard says, never how it was applied. The second is the published record of a party to the events, COGAT and two other Israeli government bodies, whose counts are their own, self-reported and not independently verified; where a publisher states its method, the method is carried with the number. The third is a humanitarian coordination body’s field enumeration of assistance received at enumerated sites, alongside a second cluster’s household survey. The fourth is one peer-reviewed qualitative study and one damage assessment co-published by the World Bank, the European Union and the United Nations.
Counts from different bodies are not summed or compared. A tonnage recorded at a crossing and a household count at a site measure different objects on different universes, so each is reported with what it covers.
Three limitations are stated rather than resolved. The report transmitted to Congress on 10 May 2024 is described and never quoted, because a search on 19 August 2026 found no copy on a host belonging to the issuing body. No determination and no waiver notice under section 620I was located, in either direction, so the section is described from its own text. And these sources reach back only to the current war, so the period before October 2023 sits outside this report.
One further absence is recorded because a reader may look for it: nothing among these sources measures how the humanitarian-access framing circulates in public discourse, so no section here addresses that question.
Conclusion
So does Israel prevent Palestinians in Gaza from accessing humanitarian aid, in the sense the United States Congress defined in 1996? The test it wrote is broad enough to reach a restriction that is indirect and short of a formal ban, and it handed the answer to a single office. That office produced its assessment in May 2024, and the assessment cannot be read on any government host. The standard is public; its application is not.
What is public is two counts of two different objects. At the crossings, a party to the events publishes tonnage compiled from what carriers declared: 1,778,363 metric tons of food across the ceasefire window, more than 106,000 trucks across the war, precise to the ton and silent on receipt. At the sites, enumerators who walked 116 neighbourhoods publish what households said they had received: 136,000 of them nothing in six months, 7,700 sleeping outdoors, 1.5 million people under roughly 246,000 tents built for five. Between those ends lies a network with more than 371,888 housing units damaged or destroyed and more than half its hospitals non-functional, and no body measures the passage end to end.
A supply chain has two ends and a middle. The record measures the ends carefully, by different hands and to different purposes, and the one document built to judge the middle went to six committees and reaches everyone else, if at all, through another party’s posting. When the answer to a question is written into law, produced on schedule and published nowhere by the body that produced it, what is a reader left to weigh the question with?
Frequently Asked Questions
Which United States statute names humanitarian access as a condition on assistance?
Section 620I of the Foreign Assistance Act, which the Law Revision Counsel publishes at 22 U.S.C. 2378-1, bars assistance under that Act or the Arms Export Control Act to a country whose government “prohibits or otherwise restricts, directly or indirectly, the transport or delivery of United States humanitarian assistance.” It carries a national security waiver.
Does the February 2024 presidential memorandum bind the countries that receive United States weapons?
It does not. National Security Memorandum 20 is a presidential directive addressed to United States officials, and the assurances it requires are the recipient government’s own written representations about its future conduct. A later President can rescind it without an act of Congress.
Why is a tonnage recorded at a crossing not a measure of what households received?
Because it counts material at the point of entry. COGAT recorded 1,778,363 metric tons of food for the ceasefire window as its own count, compiled from tonnages the carrying organisations declared, and no figure in Israel’s series follows a consignment to a household.
Who counted Gaza’s displacement sites, and how?
The Site Management Cluster deployed 165 trained enumerators from twelve partner organisations across 116 neighbourhoods between 3 February and 19 April 2026, visiting every site on its master list and confirming 1,591 of 1,856 as active.
What does the cluster mean when it calls a site makeshift?
Makeshift is a site type, not a shelter type: informal, open-air and tented, as distinct from a collective centre in a school or public building and from a scattered site of fewer than ten households. The Site Management Cluster records them as 83% of the 1,591 confirmed.
Why does the widely quoted water figure rest on 342 households?
The WASH Cluster measured the 82% water-insecurity share on a phone survey of 342 assessed households conducted between 2 and 15 March 2026, whose authors state that its findings “cannot be considered representative of the population in Gaza.” Relayed as a share of families, it states a population share the exercise did not measure.
Are the damage and needs figures in the joint assessment two versions of one number?
They are not. The World Bank assessment puts US$35.2 billion in physical infrastructure damages already incurred and US$71.4 billion in recovery and reconstruction needs over the next decade. Damages are backward-looking, needs estimate a future cost, and the two are never summed.
Has any body determined that the section 620I standard was crossed?
This report located no published determination under section 620I, and no waiver notice, in either direction. The section produces a public artefact only when the President makes a determination or files notice of an intention to waive, so silence points in neither direction.
Sources
- COGAT, 2026. Humanitarian Situation Report During the Ceasefire. Coordinator of Government Activities in the Territories, Israeli Ministry of Defense. gaza-aid-data.gov.il. Official record, a party’s own reported counts.
- Hamamra, Mahamid and Mayaleh, 2025. Childhood under siege: battling malnutrition, disease, and despair during genocide in Gaza. Journal of Health, Population and Nutrition, 44:321. DOI 10.1186/s41043-025-00948-6. Peer-reviewed, qualitative interview study.
- Israel Defense Forces, 2025. Clearing Up the Misconceptions: Everything You Need to Know About the Humanitarian Aid Situation in Gaza. idf.il, page date-stamped 19 August 2025. Official record, a party’s own reported counts.
- Israel Ministry of Foreign Affairs, 2023. Israel-Hamas Conflict 2023: Humanitarian Efforts. gov.il. Official record, a party’s own reported counts.
- Israeli Ministry of Foreign Affairs, Permanent Mission of Israel to the Multilateral Organizations in France, 2025. Israel’s Objectives and Current Operation. embassies.gov.il, last updated 17 September 2025. Official record, a party’s own reported counts.
- Site Management Cluster in the Occupied Palestinian Territory, 2026. Gaza Displacement Site Verification Assessment: Baseline Findings for May 2026 Humanitarian Planning and Coordination. cccmcluster.org, published 22 June 2026. Humanitarian coordination body, primary field enumeration.
- The White House, 2024. National Security Memorandum on Safeguards and Accountability With Respect to Transferred Defense Articles and Defense Services (NSM-20), 8 February 2024. Daily Compilation of Presidential Documents, DCPD-202400098, govinfo.gov. Primary record, presidential directive.
- UNICEF State of Palestine, 2026. Humanitarian Situation Update: Analysis of the Response to the Needs of Children at Mid-year 2026, 17 July 2026. unicef.org/sop. Agency situation update and funding appeal.
- United States Congress, 1996. Foreign Assistance Act of 1961, section 620I, codified at 22 U.S.C. 2378-1, added by Public Law 104-208, 110 Stat. 3009-121. Office of the Law Revision Counsel, uscode.house.gov. Enacted statute.
- US Department of State, 2024. Report to Congress under National Security Memorandum 20, transmitted 10 May 2024. No copy located on state.gov or on the archived 2021-2025.state.gov, searched 19 August 2026. Official record, not published by the issuing body.
- WASH Cluster, 2026. Light Touch Monitoring, Gaza, March 2026. Phone survey of 342 assessed households, 2 to 15 March 2026. Humanitarian coordination body, non-representative household survey.
- World Bank, European Union and United Nations, 2026. Gaza Strip Rapid Damage and Needs Assessment, April 2026. thedocs.worldbank.org. Joint intergovernmental assessment, co-published.