Data Reportby The Hanover InstituteSeptember 29, 2026

What Is the History Behind Claims That Israel Violates Gaza Ceasefires? Reading Seven Truces Since 2008 and the Research on Who Breaks Them

Seven Gaza truces since 2008, each ended by an incident each side blamed on the other. What the dated record and the research on lulls show.

What Is the History Behind Claims That Israel Violates Gaza Ceasefires? Reading Seven Truces Since 2008 and the Research on Who Breaks Them

Gaza has had at least seven ceasefires or pauses since June 2008 that carry a dated public record, and at the end of almost every one, each side’s own statements blamed the other. The measured research is sharper and split. Three academics found in 2009 that 79% of the lulls in lethal violence during the Second Intifada were broken first by an Israeli killing, while a peer-reviewed economics study found Israeli violence followed Palestinian violence and not the reverse. Israel’s own records attribute the collapse of each Gaza truce to Hamas.

Key Findings

  • 79% of the pauses in lethal violence from September 2000 to October 2008 ended when Israel killed a Palestinian, 8% when Palestinians killed an Israeli, and 13% with both on the same day, on an analysis by Kanwisher, Haushofer and Biletzki (2009) that was not peer-reviewed.
  • Jaeger and Paserman (2008), in the American Economic Review, found “Granger causality from Palestinian to Israeli violence, but not vice versa” in daily fatalities from September 2000 to January 2005; Haushofer and colleagues (2010), in PNAS, found that “both sides retaliate” once rocket fire is counted.
  • Of the 362 rockets and mortar shells an Israeli research centre counted during the June to December 2008 lull, 324 came after an IDF operation on 4 November 2008 that killed Hamas operatives at a cross-border tunnel (Intelligence and Terrorism Information Center, 2009).
  • Israel’s 2015 report on the 2014 conflict lists eight suspensions of hostilities and attributes the breach of each to Hamas, the party’s own account (Israel Ministry of Foreign Affairs, 2015).
  • Security Council Resolution 2803, endorsing the October 2025 arrangement, uses the word violation zero times, while the IDF’s own list records 61 dated incidents it attributes to Hamas (United Nations Security Council, 2025; Israel Defense Forces, 2026).

Which Gaza ceasefires does the dated record hold since 2008?

Seven arrangements since June 2008 carry a dated record in the sources assembled here: the lull of 19 June 2008, the understandings of 21 November 2012, the suspensions of the 2014 conflict, the ceasefire of 21 May 2021, the pause of November 2023, the ceasefire of 19 January 2025 and the arrangement in force from 10 October 2025. Only the last has a text published as an annex to a Security Council resolution.

The others took the forms their records describe. The 2008 lull, in the words of the Israeli Intelligence and Terrorism Information Center, “was based on unwritten understandings and called for the cessation of the fighting in the Gaza Strip” (Intelligence and Terrorism Information Center, 2008). The 2012 understandings were announced by Egypt’s foreign minister, with the text posted on the Egyptian president’s Facebook page (Intelligence and Terrorism Information Center, 2012). The 2021 ceasefire was an Egyptian initiative Israel’s security cabinet accepted as “mutual” and “unconditional” (CNN, 2021).

Source: ITIC, 2008 and 2012; Israel Ministry of Foreign Affairs, 2015; CNN, 2021; CBS News, 2023 and 2025; Security Council resolutions 1860 (2009) and 2803 (2025).

The list is not exhaustive. It leaves out shorter truces, including the August 2022 ceasefire with Palestinian Islamic Jihad, for which no primary record was read for this report. What the seven share is the absence, until 2025, of any written text defining what counts as a breach. Every judgment about who broke an arrangement before 2025 was therefore a judgment made by one of the parties, a monitor or a researcher, never one set by the arrangement itself. The sections below take the research first, then each arrangement in turn.

Is it often said Israel repeatedly broke Gaza ceasefires?

It is often said that Israel has repeatedly violated ceasefires in Gaza, and the claim has a dated quantitative source. In February 2009, Nancy Kanwisher, Johannes Haushofer and Anat Biletzki reported that 79% of all conflict pauses in the Second Intifada “were interrupted when Israel killed a Palestinian, while only 8% were interrupted by Palestinian attacks,” the remaining 13% by both sides on the same day (Kanwisher, Haushofer and Biletzki, 2009).

Their method is stated in the piece. They took the daily record of killings on both sides from September 2000 to October 2008, compiled by the Israeli human rights organisation B’Tselem, defined a “conflict pause” as one or more days on which no one on either side was killed, and asked which side killed first when each pause ended. The piece was published on the Huffington Post, not in a journal, and it counts killings only: a pause ended by rocket fire that killed no one would not register as broken.

The same authors describe the 2008 lull in terms Israel’s own records share. After it began in June 2008, “the rate of rocket and mortar fire from Gaza dropped to almost zero, and stayed there for four straight months,” a pattern they illustrate with a factsheet from the Israeli consulate in New York. On 4 November, they write, “Israel killed a Palestinian, an event that was followed by a volley of mortars fired from Gaza,” and a further strike killed six more.

The claim is supported in the form its authors measured: in their data, the first killing after a lull was far more often Israeli than Palestinian. It is not a finding about rocket fire, about intent, or about the arrangements after 2008, which their data do not cover. The peer-reviewed literature, taken up next, disputes what the ordering of violence shows.

What does peer-reviewed research find about who breaks the lulls?

Two peer-reviewed studies of the same conflict reach opposite readings. David Jaeger and Daniele Paserman, in the American Economic Review in 2008, analysed daily fatalities from September 2000 to January 2005 and found “evidence of Granger causality from Palestinian to Israeli violence, but not vice versa,” with “no evidence that the Palestinians and Israelis are engaged in a predictable ‘tit-for-tat’ cycle of violence” (Jaeger and Paserman, 2008).

Johannes Haushofer, Anat Biletzki and Nancy Kanwisher, in the Proceedings of the National Academy of Sciences in 2010, replicated the first half of that result: “Israeli killings of Palestinians increase after Palestinian killings of Israelis.” They then widened the data, adding rocket launches from IDF records alongside B’Tselem fatality counts, and found that “the firing of Palestinian rockets increases sharply after Israelis kill Palestinians” and that “the probability (although not the number) of killings of Israelis by Palestinians increases after killings of Palestinians by Israel” (Haushofer, Biletzki and Kanwisher, 2010).

The two findings differ in what they count. Jaeger and Paserman count fatalities; Haushofer and colleagues add non-lethal rocket fire and a longer series running from 2000 to 2008. Their title states the conclusion: “Both sides retaliate.” The PNAS study drew a published critique from Golan and Rosenblatt (2011), and the authors’ reply in the same journal is titled “Revisiting the statistical analysis of the Israeli–Palestinian conflict confirms that both sides retaliate” (Haushofer, Biletzki and Kanwisher, 2011).

Granger causality is a statistical test of whether one series helps predict another; it measures ordering in time, not intent or legal responsibility. Neither study addresses who breached a particular agreement, and neither covers any arrangement after 2008. What they establish is that the answer to “who responds to whom” depends on which forms of violence are counted, which is why each side’s account of any single truce can be internally consistent and still contradict the other’s.

How did the June 2008 lull end, on Israel’s own count?

324 of the 362 rockets and mortar shells fired at Israel during the June to December 2008 lull came after 4 November, on the count of the Intelligence and Terrorism Information Center, an Israeli research institute (Intelligence and Terrorism Information Center, 2009).

From 19 June to 4 November, the centre’s own count records 20 rocket landings and 18 mortar shells. Its report of 19 December 2008 gives Israel’s account of the turning point: “On November 4 the IDF carried out a military action close to the border security fence on the Gazan side to prevent an abduction planned by Hamas, which had dug a tunnel under the fence to that purpose. Seven Hamas terrorist operatives were killed during the action. In retaliation, Hamas and the other terrorist organizations attacked Israel with a massive barrage of rockets” (Intelligence and Terrorism Information Center, 2008).

Source: Intelligence and Terrorism Information Center, 2009, Summary of rocket fire and mortar shelling in 2008; the centre’s own count.

The centre’s figures carry their own context. It records that the lull “brought relative quiet to the western Negev population and the Gaza Strip, especially during its first months,” and that the six months before it saw 2,278 rockets and mortar shells, an average of 380 a month. It also records what the lull was meant to deliver: “the opening of the crossings between the Gaza Strip and Israel,” negotiations for the release of a captive Israeli soldier, and talks on Rafah.

On the timing, Israel’s own record and the 2009 analysis agree: the lull held until 4 November, and the IDF operation of that day came first. They disagree on its character. The Israeli centre calls it a preventive action against a tunnel, and the analysts count it as the killing that ended a pause. Neither description is a finding by a body appointed to judge the arrangement, because the arrangement, being unwritten, appointed none.

What did the Security Council ask for in January 2009?

14 of the Security Council’s 15 members voted on 8 January 2009 for Resolution 1860, with the United States abstaining and none against, according to the verbatim record of its 6063rd meeting (United Nations Security Council, 2009). Operative paragraph 1 “calls for an immediate, durable and fully respected ceasefire, leading to the full withdrawal of Israeli forces from Gaza.”

The resolution’s preamble describes the collapse of the 2008 lull without assigning it to either side. It expresses “grave concern at the escalation of violence and the deterioration of the situation, in particular the resulting heavy civilian casualties since the refusal to extend the period of calm.” The phrase names a refusal and no refuser. The same preamble emphasises “the need to ensure sustained and regular flow of goods and people through the Gaza crossings,” the issue the lull had been meant to address.

The resolution sets no mechanism for determining a breach. It “Condemns all violence and hostilities directed against civilians and all acts of terrorism,” calls for the unimpeded provision of humanitarian assistance, and welcomes initiatives for humanitarian corridors. It names no monitor, no verification body and no consequence for non-compliance.

The Council’s text therefore did what every arrangement before 2025 did: it called for a ceasefire whose terms and whose breaches were left to the parties and the mediators. The one party-neutral document of the 2008 to 2009 period is a call for a ceasefire, not a finding about the one that ended. The next arrangement with a dated record came almost four years later, through Egyptian mediation.

What did the November 2012 understandings require?

The 2012 understandings came into effect at 21:00 on 21 November 2012, brokered by Egypt with American involvement, according to the Intelligence and Terrorism Information Center’s update of that day. It reports that the text, posted on the Egyptian president’s Facebook page, provided that “the two sides agreed to stop all offensive activities and within 24 hours deliberations would begin about broader issues” (Intelligence and Terrorism Information Center, 2012).

The centre records that after 21:00, “ten rocket hits were identified in Israeli territory, but since 2300 hours there has been quiet.” Israel, Hamas and Palestinian Islamic Jihad each announced they accepted the ceasefire. Israel’s prime minister announced that the operation had ended and listed what he described as its achievements; the centre relays these from the prime minister’s website.

The understandings were announced by a mediator rather than signed by the parties. The Egyptian foreign minister presented them at a joint press conference with the United States Secretary of State. No text was deposited with the United Nations or published by either party as an agreement. The broader issues that were to be deliberated within 24 hours, which the centre does not itemise, have no published outcome in the sources assembled here.

This is the pattern the record shows through 2021: a mediator announces a mutual cessation, each side accepts it on its own terms, and the question of what either side must do beyond stopping fire is deferred. When fighting resumed in 2014, there was no written standard against which either side’s conduct under the 2012 understandings could be measured, and the sources assembled here record no determination of who broke them.

How does Israel’s record describe the 2014 truces?

Eight suspensions of hostilities appear in the table in Israel’s May 2015 report on the 2014 Gaza conflict, which calls the list “non-exhaustive” (Israel Ministry of Foreign Affairs, 2015). The report states that “Israel agreed to numerous ceasefires brokered by other countries and international organisations” and “also declared unilateral ceasefires.”

The table records each suspension with its sponsor and, in the report’s own column headed “Violations,” what ended it. The Egyptian proposal of 15 July was “Rejected by Hamas; 56 rockets fired.” A suspension proposed by the Security Council for 28 July was followed, on the report’s account, by “Continued firing of rockets into Israel, infiltration through a tunnel into Israel, and attacks against IDF forces.” The three-day suspension from 08:00 on 1 August, coordinated on a UN and US proposal, was cancelled after an “Attack against IDF forces, attempted abduction of IDF soldier.” The five-day suspension from 14 August ended with “50 Rockets and mortars were fired towards Southern Israel after the suspension was extended by 24 hours.”

The report states its conclusion at paragraph 396: Hamas “consistently rejected proposed ceasefire agreements, violated coordinated ceasefires, and exploited unilateral IDF ceasefires by conducting military activities against the IDF and carrying out rocket and mortar attacks against Israel.” It records the use Israel made of one pause: during the suspension of 26 and 27 July, repairs raised the electricity Israel supplied to Gaza “from 52mw to 102mw.”

This is the party’s own record of its own conduct and the other side’s, published by the government of a state that fought the conflict; it lists no breach by Israel, and no independent body is cited as having verified its entries. It is nonetheless the most detailed dated list of the 2014 suspensions in the sources assembled here, and it shows a structure the other arrangements share: each pause was proposed by a third party, accepted or rejected by each side separately, and ended by an incident one side attributed to the other.

How did the 2021 ceasefire and the 2023 pause end?

The 2021 ceasefire was set to take effect at 2 a.m. on 21 May 2021, a senior Hamas official told CNN, after Israel’s security cabinet accepted “the Egyptian initiative for a mutual unconditional ceasefire,” as a statement from the prime minister’s office put it (CNN, 2021). The cabinet acted, the statement said, on the recommendation of the Chief of Staff and the heads of the Shin Bet, the Mossad and the National Security Council.

The word “unconditional” carries the point. The 2021 arrangement, like 2012, was a mutual cessation announced through a mediator with no published terms beyond stopping fire, which is why its end cannot be measured against a text. The sources read for this report record no published determination of how it later broke down.

The pause of November 2023 was different in kind, because it was tied to hostage releases. It began on 24 November 2023 and ran for a week; CBS News reported that 105 hostages were freed during it, along with some 240 Palestinian prisoners, the latter figure from the Palestinian Prisoners’ Club (CBS News, 2023). As it was due to expire at 7 a.m. on 1 December, the Israeli military said it had intercepted a rocket fired from Gaza, and then posted: “Hamas violated the operational pause and in addition, fired toward Israeli territory. The IDF has resumed combat against the Hamas terrorist organization in the Gaza Strip.”

That statement is the party’s own account, and it is the only account of the pause’s end in the sources read for this report; no mediator’s determination of which side ended it was located. What distinguished 2023 from the earlier pauses is that it had a countable exchange at its centre. Its end therefore turned not only on fire but on whether the next release would happen, a question the January 2025 ceasefire made explicit in phases.

How did the January 2025 ceasefire end?

The January 2025 ceasefire ran from 19 January, and its first phase ended on 1 March 2025; the pause in major fighting held until Israel resumed strikes on 18 March, CBS News reported (CBS News, 2025). The two sides could not agree on a second phase aimed at releasing the remaining hostages and ending the war.

Each side’s account of the end is on the record, and they are opposite. The statement from the Israeli prime minister’s office said Israel was resuming strikes because of Hamas’s “repeated refusals to release its hostages and its rejection of all offers it received from the U.S. presidential envoy Steve Witkoff and mediators,” as CBS reported it. Hamas, in a statement quoted in the same report, said that “Netanyahu and his extremist government have decided to overturn the ceasefire agreement,” and that it held the Israeli government “fully responsible for the repercussions.”

The report records the context both statements sit in. The first phase had seen dozens of hostages released in exchange for nearly 2,000 Palestinian prisoners. After it ended, the report says, Israel’s prime minister had repeatedly threatened to resume the war, and the United States president had issued warnings and ultimatums to Hamas, “some contradicting the agreed-upon terms of the ceasefire.” Egypt’s foreign ministry called for immediate international intervention to halt the strikes.

This is the first arrangement in the record whose breakdown turned on the terms of a phased agreement rather than on a first shot. Each side cited the other’s refusal: Israel of further releases on the proposals put to Hamas, Hamas of the agreement’s second phase. No mediator’s published determination of which reading was right appears in the sources read for this report, and the text of the January 2025 agreement itself was not among them.

What does the October 2025 arrangement count as a violation?

Zero times: Security Council Resolution 2803, adopted on 17 November 2025 by 13 votes to none with China and the Russian Federation abstaining, never uses the word violation, and neither does the twenty-point plan annexed to it (United Nations Security Council, 2025). The plan says “the war will immediately end” and that military operations “will be suspended” and battle lines “frozen.”

The only monitoring clause is thin. Operative paragraph 7 has the International Stabilization Force “assist the BoP in monitoring the implementation of the ceasefire in Gaza,” and paragraph 10 asks the Board of Peace for “a written report on progress” every six months. The Board’s only report, of 15 May 2026, states that “violations continue to occur on a near daily basis, some of which are serious,” with no total, rate or attribution to a party (Board of Peace, 2026).

The parties fill the gap with their own lists. The Israel Defense Forces publish a page, dated 19 February 2026, compiling what they present as “IDF press releases addressing each ceasefire violation by Hamas since the start of the agreement,” with 61 dated entries from 19 October 2025 to 14 February 2026 (Israel Defense Forces, 2026). The page states no total, applies the IDF’s own test that the activities “constitute a clear violation of the understandings of the ceasefire,” and records no incident by Israel. Palestinian bodies publish counts in the other direction; they reach readers through outlets this report does not rely on and are not reproduced here.

The 2025 arrangement is thus the first with a published text, and the text still leaves the question this report asks where every earlier arrangement left it: with the parties. The Institute’s report on what the ceasefire instrument defines reads the 2025 texts in detail.

Methodology and limitations

This report draws on five types of source. Primary instruments and records are Security Council resolutions 1860 (2009) and 2803 (2025) with their recorded votes. Party records are Israel’s May 2015 report on the 2014 conflict, the IDF’s 2026 compilation, and statements from Israel’s prime minister’s office and military as relayed by named press. Israeli research: the Intelligence and Terrorism Information Center’s reports of 2008, 2009 and 2012, which are the centre’s own counts and characterisations. Peer-reviewed research: Jaeger and Paserman (2008), Haushofer and colleagues (2010, 2011) and the Golan and Rosenblatt (2011) critique. Press and commentary: the 2009 analysis by Kanwisher, Haushofer and Biletzki, which was not peer-reviewed, and CNN and CBS News reports that relay dated statements.

The research measures ordering in time, not legal responsibility. The 2009 analysis counts only killings and relies on B’Tselem data, a monitoring organisation’s own compilation; the PNAS study adds IDF rocket data; the AER study covers 2000 to 2005 only. None covers any arrangement after 2008.

The arrangement-by-arrangement record is uneven. Before 2025, no arrangement had a published text defining a breach, so every account of an ending is a party’s, a mediator’s or a researcher’s. The texts of the 2012 understandings as posted and of the January 2025 agreement were not read directly, and the August 2022 ceasefire is not covered. Israeli statements are carried in full where a party’s own account is the only one available, and labelled as such.

Conclusion

So what is the historical context behind the claim that Israel violates ceasefires in Gaza? It is a record of at least seven arrangements since 2008, almost all unwritten, each ended by an incident that each side attributed to the other, and a research literature that agrees on the ordering of events far less than either side’s account suggests.

The record is specific where it can be. The 2008 lull held until an IDF operation on 4 November, and 324 of its 362 projectiles came after it, on Israel’s own centre’s count. A 2009 analysis found Israel killed first after 79% of Second Intifada lulls; a peer-reviewed study found Israeli violence followed Palestinian violence, and another found that both sides retaliate once rockets are counted. Israel’s 2015 report lists eight 2014 suspensions and attributes every breach to Hamas. In 2025 each government cited the other’s refusal. And the one arrangement with a published text never uses the word violation.

The pattern is less a record of who broke each truce than of who was never asked to decide. When no arrangement names a judge, every count of violations is someone’s own, and the question becomes which count a reader already trusts before reading it. If the next ceasefire again leaves breaches undefined, whose list will the history of its ending be written from?

Frequently Asked Questions

Was the 2008 lull written down?

No. The Intelligence and Terrorism Information Center describes it as “based on unwritten understandings” that “called for the cessation of the fighting in the Gaza Strip,” with Hamas committing to enforce it on other groups (Intelligence and Terrorism Information Center, 2008). It was expected to last six months, and Hamas spokesmen said it would end on 19 December 2008.

Did the Security Council blame either side for the end of the 2008 lull?

No. Resolution 1860 refers to casualties “since the refusal to extend the period of calm” without naming who refused, and calls for “an immediate, durable and fully respected ceasefire” (United Nations Security Council, 2009). It was adopted by 14 votes to none, with the United States abstaining.

How many rockets were fired before the 2008 lull, on Israel’s count?

The Intelligence and Terrorism Information Center counted 2,278 rockets and mortar shells in the six months before the lull, an average of 380 a month (Intelligence and Terrorism Information Center, 2008). During the six-month lull it counted 329 by 17 December 2008, most of them after 4 November.

What is Granger causality, and why does it matter here?

It is a statistical test of whether past values of one series help predict another. Jaeger and Paserman (2008) found Palestinian fatalities predicted Israeli ones and not the reverse; it measures timing, not intent or who breached an agreement.

Did Israel supply more electricity to Gaza during the 2014 pauses?

Israel’s 2015 report states that during the suspension of 26 and 27 July 2014, repairs to power lines raised the electricity Israel supplied to Gaza “from 52mw to 102mw” (Israel Ministry of Foreign Affairs, 2015). That is the party’s own account.

Who monitors the October 2025 ceasefire?

Resolution 2803 tasks the International Stabilization Force with assisting the Board of Peace “in monitoring the implementation of the ceasefire,” but sets no duty to count or publish violations (United Nations Security Council, 2025). The Board’s May 2026 report gives no violation total (Board of Peace, 2026).

Sources

Related research