Who was responsible for the 1948 refugees? The debate, explained.

Benny Morris’s archival history of 1948, published in its second edition by Cambridge University Press in 2004, puts the number of Palestinians who became refugees at roughly 700,000. It finds no single cause behind their departure: military attack and fear of attack, expulsions in certain sectors, and flight, varying by locality and phase. The question underneath the argument is plain: who made them leave, and who answers for what came after? On the first half, Mahmoud Yazbak told a 2016 Hebrew University workshop that military operations resulted in the expulsion and flight of about 80% of residents, and Ilan Pappé (2006) argues for a planned ethnic cleansing. On the second half, the argument turns on statutes the Knesset enacted from 1950 and on a United Nations valuation of £P100,383,784 for the land left behind.

Key Findings

• Roughly 700,000 Palestinians became refugees in 1948 on Benny Morris’s own archival estimate, drawn from Israeli military and state archives. Morris attributes the exodus to several causes that vary by locality and phase, and he rejects both a single master expulsion plan and the claim that departures were voluntary (Morris, 2004).
• About 80% of residents of the regions on which Israel was established were subject to “expulsion and flight” in the course of Zionist and Israeli military operations, on Mahmoud Yazbak’s account. Yazbak also counts about 400 towns and villages destroyed. His figures come from a rapporteur’s summary of a 2016 workshop presentation (Israel Law Review, 2018).
• 500,000 refugees was the figure Israel maintained at the time, as Morris summarised it in 2016. This was Israel’s own count as a party to the events. Arab countries, also parties, claimed nearer 900,000 or 1 million (Israel Law Review, 2018).
• Approximately 711,000 “refugees from Israel-controlled territory” was the estimate of the UN Conciliation Commission’s Technical Committee on 7 September 1949. The count names no cause of departure (United Nations Conciliation Commission for Palestine, 1950).
• Some 1.2 million dunams were expropriated in the first year of the 1953 Land Acquisition Law. Of the 311,000 dunams taken from private ownership, 304,700 came from Arab owners (Forman and Kedar, 2004).

Where does the argument over Israel’s responsibility actually divide?

Mahmoud Yazbak presented at a Hebrew University of Jerusalem workshop in December 2016, summarised in the Israel Law Review in 2018. He attributed the expulsion and flight of about 80% of residents of the regions on which Israel was established to “Zionist and Israeli military operations.” Benny Morris (2004), who spoke in the same session, reads the war as several causes that varied by place and by phase.

The two presentations sit in the first session of a two-day workshop on the 1948 refugees, held 14 to 15 December 2016 at the Faculty of Law. The published record is a rapporteur’s summary of each speaker, written in the third person. It is workshop proceedings, not a peer-reviewed research article, so each figure belongs to its speaker and not to the journal. Yazbak’s summary joins the 80% to a second count, “about 400 Palestinian towns and villages destroyed,” and assigns both to military operations.

Morris’s archival history, The Birth of the Palestinian Refugee Problem Revisited (Cambridge University Press, 2004), lists the causes: military attack and fear of attack, expulsions in certain sectors, and flight. It finds that the mix changed from locality to locality and from one phase of the war to the next. Morris rejects two positions at once: that a single master expulsion plan existed, and that the departures were voluntary.

Set side by side, the two accounts agree that expulsions happened. Morris records expulsions in certain sectors, and Yazbak names expulsion alongside flight. They divide over weight and design: whether military operations account for most of the departures, and whether those operations followed a plan. The locality-level detail behind that dispute is laid out in a separate reading of the 1948 displacement.

A second argument runs alongside the first and stands apart from how the first is settled. It concerns what Israel enacted after the fighting, from the Absentee Property Law of March 1950 to the land legislation of July 1960. It also concerns what the United Nations Conciliation Commission for Palestine estimated the refugees had left behind and what was later released or compensated. The first argument is about the months of 1948. The second is about a statute book and a ledger that can be checked line by line.

Who says Israel bears primary responsibility, and since when?

Ilan Pappé’s The Ethnic Cleansing of Palestine (Oneworld, 2006) states the strongest scholarly form of the primary-responsibility case: the 1948 displacement was a planned ethnic cleansing. Mahmoud Yazbak’s 2016 workshop presentation, as summarised in the Israel Law Review (2018), attaches to the same events a count of about 400 Palestinian towns and villages destroyed in Zionist and Israeli military operations.

Pappé’s book is interpretive scholarship, and its author states a moral position. Its title is the named home of the ethnic-cleansing characterization of 1948. A journal article Pappé published in 2006 in the Journal of Palestine Studies makes the same argument and circulates in reprints. The thesis is about design: it treats the displacement as a planned outcome rather than a by-product of war. Ethnic cleansing is a legal and moral characterization. Here it is the dated thesis of a named author, not the finding of any court.

Yazbak’s 2016 presentation brings a reading close to Pappé’s into a joint session with Morris, whose archival history is the counter-pole to Pappé’s thesis.

Morris (2004), drawing on Israeli military and state archives, lists expulsions in certain sectors among the causes of the exodus. Expulsion therefore appears in both accounts. Planning is where they part: Pappé argues the displacement was planned, and Morris rejects a single master expulsion plan. The dispute is between named historians, and neither book settles it.

Yazbak’s two figures carry a date and a speaker: they come from a 2016 presentation and the rapporteur’s 2018 summary of it. They also measure different things from the refugee counts in later sections. The roughly 400 villages counts places destroyed, and the 80% is a share of residents of a defined region. Neither can be added to, or set against, the UN Technical Committee’s estimate of people. The responsibility case on this side rests on cause and design, carried by Pappé’s thesis and Yazbak’s counts, rather than on the size of the refugee population.

What does the Israeli archival record say caused the departures?

Benny Morris’s The Birth of the Palestinian Refugee Problem Revisited (Cambridge University Press, 2004) is built from Israeli military and state archives. It gives roughly 700,000 as its own estimate of the Palestinians displaced in 1948. It attributes their departure to military attack and fear of attack, expulsions in certain sectors, and flight, in proportions that vary by locality and by phase of the war.

The book is the second edition of an archival history, published in the Cambridge Middle East Studies series. It is interpretive scholarship, not a measurement. It supports a typology of causes and an overall estimate, not a rate of expulsion that could be compared across studies. Its value to the responsibility argument lies in its sources: it reconstructs events from records kept by the Israeli state and its armed forces.

On those records, Morris supports neither pole whole. He rejects the claim that the departures were voluntary. He also rejects a single master expulsion plan, which sets him against Pappé’s 2006 thesis of a planned displacement. What remains is a locality-by-locality account in which some departures followed attack or the fear of it, some followed expulsion orders in specific sectors, and some were flight.

Morris’s account also fixes where the disagreement can be checked. Because the typology varies by place and phase, the dispute between Morris and Pappé is partly over aggregation. One question is whether the expulsions Morris documents in certain sectors were local decisions or parts of a design. The other is how much of the total each cause accounts for. Morris (2004) answers with multiplicity and Pappé (2006) with planning. Both books are named, dated and published by established presses, and the scholarly dispute between them stands as a dispute.

The book is also distinct from the Israeli government’s position. Morris’s roughly 700,000 is a historian’s later estimate built from the archives. The figure the Israeli government put forward at the time, carried in the next section, was lower. The gap between a state’s contemporary count and a later history written from that state’s own files is itself part of the record on responsibility.

How many refugees did Israel, Arab governments and the UN each claim?

Israel maintained at the time that the number of refugees was 500,000. This was its own count as a party to the events, not an independently verified figure, and it comes from Benny Morris’s 2016 workshop presentation as summarised in the Israel Law Review (2018). It is the lowest of the three contemporary claims Morris set side by side.

In the same summary, Morris states that Arab countries, also parties to the war, claimed a figure “nearer 900,000 or one million.” The United Nations, in 1949 to 1950, concluded that the real figure was “just over 700,000.” The contemporary claims therefore span a factor of two, from Israel’s 500,000 to the 1 million at the upper end of the Arab claim.

Each figure belongs to the body that made it, and none is averaged against another. The two party figures are claims made while the question of responsibility was live. The UN figure is an estimate by a technical body, set out in the next section. Morris’s roughly 700,000 (2004) is a later archival estimate and lands close to the UN number, but it is a separate exercise.

The size of the problem was contested in its own time. A count of refugees carries an implied scale of return or compensation, so the parties’ figures were part of the responsibility argument from the start. The spread also has a documentary basis that the UN Technical Committee itself recorded: relief rolls that ran higher than its estimate, for reasons it set out.

One further number circulates widely. A figure of 726,000 belongs to the UN Economic Survey Mission, a different body that reported separately from the Conciliation Commission. That figure does not appear in the Commission’s 1950 report, and it is a separate series from the 711,000.

Did the UN’s 1949 refugee count say why people left?

The Technical Committee on Refugees of the UN Conciliation Commission for Palestine estimated approximately 711,000 “refugees from Israel-controlled territory” in a report dated 7 September 1949.

The report was published in the Commission’s 1950 progress report (A/1367/Rev.1, Appendix 4, paragraph 15). The phrase counts people by where they came from, and the document gives no reason for any departure.

The Committee’s vocabulary is consistent across the appendix. It writes of refugees “from” territory, and nowhere says anyone was expelled, driven out or left voluntarily. Both sides of the responsibility argument cite UN figures. This one is a count without a cause, so it cannot carry either position about why people left. For cause, the record turns to the archival and interpretive histories by Morris (2004) and Pappé (2006).

The figure is an estimate, not an enumeration. The Committee’s authors label it only “as accurate as circumstances permit.” The estimate was prepared by a statistical expert for the Committee in Lausanne. The Committee considered taking a census and concluded that one was not advisable. Paragraph 11 explains part of the reasoning: an estimated 500,000 residents of the former Mandatory area were apparently self-supporting, and a census risked attracting new claims to refugee status from among them.

Paragraph 15 accounts for why the relief rolls carried a higher number than 711,000. It names three sources of the excess: duplication of ration cards, persons displaced from areas other than Israel-held areas, and persons who were destitute without having been displaced. The Committee treated the rolls as an overcount of the population it was estimating and arrived at its figure by other means.

The dates matter for citation. The 711,000 dates from 7 September 1949, the report that carries it is dated 23 October 1950, and the printed volume appeared in 1951. The accurate form is the Commission’s 1949 estimate, published in its 1950 report. That is how the figure appears in the chart above, beside the parties’ own claims and Morris’s later archival estimate.

Which Israeli laws figure in the continuing-responsibility argument?

The Knesset’s 1953 Land Acquisition Law expropriated some 1.2 million dunams in its first year of operation, according to Forman and Kedar’s peer-reviewed legal history in Environment and Planning D (2004). The authors place the statute within a 12-year sequence that brought the land of Palestinians displaced in 1948 into state hands.

Geremy Forman and Alexandre Kedar, both of the University of Haifa, built their account from Israeli legislation, archival documents and Knesset proceedings. They title the process “legal dispossession.” They set out four phases between 1948 and 1960, each closing with a major statute. The core mechanism, in their words: “The Knesset enacted the Absentee Property Law in March 1950 and the Development Authority (Transfer of Property) Law in July 1950.”

The two laws set a single path for the land. “The Absentee Property Law empowered the CAP to sell land to the DA alone, and the Development Authority Law empowered the DA to sell land to the state and the JNF.” CAP is the Custodian of Absentee Property, and DA is the Development Authority. The authors describe the Development Authority as an intermediary body, legally defined as extragovernmental, that passed expropriated land to the state and the Jewish National Fund. A separate reading of the absentee-property texts follows the claims filed under them.

The 1953 figures break down as subsets of the 1.2 million, not additions to it. Some 311,000 dunams were taken from private ownership, 304,700 of those from Arab owners, and 704,000 dunams were taken from the Custodian. The authors state that it is not clear how much of the 704,000 was land of Arab citizens classified as absentees while still resident in Israel.

The sequence closes in late July 1960. That month the Knesset enacted Basic Law: Israel Lands, the Israel Lands Law and the Israel Lands Administration Law. Together they consolidated state, Development Authority and Jewish National Fund holdings into one category under a single administration. Forman and Kedar’s account ends there and covers no later statute.

What value did the UN put on abandoned refugee land?

In September 1951, the Refugee Office of the UN Conciliation Commission for Palestine valued the abandoned land that had passed into Jewish hands at £P100,383,784.

The Commission reported that estimate in a working paper dated 2 October 1961 (A/AC.25/W/81/Rev.2, Annex V). The paper addresses the compensation question under paragraph 11 of General Assembly resolution 194 (III).

The same annex gives the extent: 16,324 square kilometres of abandoned land, of which 4,574 square kilometres were cultivable. The total value breaks into three parts. Rural lands were valued at £P69,525,144, urban lands at £P21,608,640 and Jerusalem lands at £P9,250,000. All values are in Palestine pounds as at 29 November 1947.

The method is stated in the working paper. For the value, the Refugee Office combined the Mandatory government’s property-tax assessments with expert opinion. It treated buildings and trees as part of the soil on which they stood. The Office attached its own warning: the estimates “must inevitably be regarded as approximate ones.”

The valuation is the Commission’s own estimate on its own definitions, and it measures abandoned land. It does not measure land that moved through any particular Israeli statute, a distinction the Office drew deliberately, as the last section sets out. The estimate was made in 1951 and reported in 1961, and the two dates refer to different things.

In the responsibility argument, the figure gives the compensation side a dated intergovernmental number. Here, the claim for compensation is tied to an estimate by a UN body with a stated method. The working paper presents it as an approximate valuation for the purpose of the compensation question, and the Commission did not treat it as a settled liability.

How much refugee property had been released or compensated by 1959?

Compensation under the 1953 land-acquisition mechanism had been paid for only 64,500 dunams by 1959. Forman and Kedar’s peer-reviewed 2004 article relays the figure from a 1959 account by Liskovski. It is the earliest of the compensation totals the article relays for that mechanism.

The figure covers one statute and one kind of payment. It records dunams compensated under the 1953 Land Acquisition Law, the law that, on the same authors’ account, expropriated some 1.2 million dunams in its first year. The 64,500 counts land for which payment was made, not money paid. It is not a share of any total, because the article gives no figure for how much of the expropriated land was eligible for compensation under that law.

A second, separate fragment of restitution appears in the UN Conciliation Commission’s 1961 working paper. Release of Arab refugee bank accounts blocked in Israel began in the summer of 1953 at the Palestine branches of Barclay’s Bank and the Ottoman Bank. The Commission records that “By 31 August 1956 a total of £2,633,175 of the accounts had been released.”

The two figures measure different things and do not combine. The £2,633,175 is a released balance: depositors’ own money in blocked accounts returned to them. It is not compensation for property, so it cannot be set against the Refugee Office’s £P100,383,784 valuation of abandoned land as a partial payment toward it. The 64,500 dunams is a count of land, not of money, and it cannot be converted into either sum.

Within the 1950s, then, the record holds three numbers on the property question, each produced by a different process. They are a September 1951 valuation of abandoned land, a cumulative released balance on blocked accounts at 31 August 1956, and a compensated area under one statute by 1959. The Commission’s 1961 paper and Forman and Kedar’s 2004 article each report their own figures on their own definitions. No document assembled here converts any one of them into another.

Why is there no agreed figure for the land at stake?

Forman and Kedar’s 2004 legal history relays five separate estimates of the land that passed from Arab hands to the state after 1948. The authors state plainly: “There is no one agreed-upon figure for the land that moved from Arab hands to the state in the wake of 1948.” They attribute each estimate to its own source and adopt none of them.

The estimates differ partly because they count different objects. The 1951 figure is the Refugee Office’s 16,324 square kilometres of abandoned land, which is about 16.3 million dunams. The 1964 programme counted privately held land and excluded most of Beersheba. The 1964 estimate is therefore a different measurement, not a revision of the 1951 one, and neither replaces the other.

The 1961 working paper also records why the Commission’s figure was built as it was. Of four approaches weighed for the extent, the Refugee Office rejected the records of the Custodian of Absentee Property because “it was felt that it would be inappropriate to assess compensation entirely on the basis of information furnished by an interested party.” It built the estimate instead from the Mandatory government’s Village Statistics 1945.

The Office’s choice also explains a gap in the record. Because the Office set aside the Custodian’s files, the Commission’s figures measure land abandoned, not land transferred through the absentee-property system. Forman and Kedar (2004) give no total for that second object either. Neither the 1961 working paper nor the 2004 article publishes a total for land that passed through the Custodian.

The disagreement over scale tracks the disagreement over responsibility. The estimates come from a UN technical office, a later UN programme, a named researcher, and two groups the authors identify by affiliation. The bodies sit at different distances from the parties, and each counts with its own sources and its own definitions.

Methodology and limitations

The evidence for this report comes from three source types. Primary and official records: the UN Conciliation Commission for Palestine’s 1950 progress report, whose Appendix 4 carries the Technical Committee’s refugee estimate of 7 September 1949, and its 1961 working paper on the compensation question, whose Annex V carries the Refugee Office’s September 1951 land valuation. Peer-reviewed research: Forman and Kedar’s 2004 legal history in Environment and Planning D, built from Israeli legislation, archival documents and Knesset proceedings, covering 1948 to 1960. Scholarly books: Morris (2004), an archival history from Israeli military and state archives, and Pappé (2006), an interpretive thesis whose author states a moral position.

The Israel Law Review item (2018) is workshop proceedings: a rapporteur’s third-person summary of presentations given on 14 to 15 December 2016. It is not peer-reviewed. Every figure from it is attributed to its speaker, Morris or Yazbak, and dated to the workshop and the summary.

The UN refugee figure is an estimate its authors call only “as accurate as circumstances permit,” not a census, and it excludes people displaced inside Israel. The land valuation is described by the Refugee Office as approximate and is reported here in its exact form throughout. The contemporary counts by Israel and by Arab governments are each party’s own claim. Several land figures in Forman and Kedar (2004) are relayed from other works: the 1959 compensation total, which the authors relay from an account by Liskovski, and the five land estimates. Each is attributed to Forman and Kedar as relaying it, with its underlying holder named.

The central scholarly dispute, over the weight of causes and whether a plan existed, is carried here as a named disagreement between Morris (2004) and Pappé (2006). The two histories reach different conclusions about design.

Conclusion

So who made them leave, and who answers for what came after? On the first half, the record answers with a dispute between named historians rather than with a single cause. On the second, it answers with enacted statutes and a UN valuation of the land the refugees left.

The portrait is consistent across every axis measured. Morris’s archive-based history finds roughly 700,000 refugees and a mix of attack, fear, expulsion in certain sectors and flight, while rejecting both a master plan and voluntary departure. Yazbak counts about 80% of residents subject to expulsion and flight, and Pappé argues for a planned design. The contemporary counts ran from 500,000, Israel’s own figure as a party to the events, to 1 million at the top of the Arab claim. The UN Technical Committee’s 711,000 names no cause at all. After the war, the Knesset built a statutory path for absentee land that expropriated some 1.2 million dunams in one year of one law. The UN Refugee Office valued the abandoned land at £P100,383,784, and compensation under the 1953 mechanism reached 64,500 dunams by 1959.

The weight of the argument therefore sits in two places. The 1948 causes remain a scholarly dispute over aggregation and intent. The post-war record is documentary, dated and specific. A society deciding what responsibility means here faces a question that is partly historical and partly legal. If the count of people records no cause, and the statute book records every transfer, which record does a claim of responsibility rest on: the months of flight, or the twelve years of law that followed?

Frequently Asked Questions

What is a “present absentee” under Israel’s land laws?

Forman and Kedar (2004) describe the category as a result of how the absentee definition operated: “practically no Jewish Israelis, but tens of thousands of Arab Israeli citizens, were classified as absentees, assuming the paradoxical legal identity of ‘present absentee’.” These were citizens who remained inside Israel but whose property fell under the absentee regime.

From what date did someone count as an absentee?

The definition ran from 29 November 1947, the date of the UN General Assembly’s partition resolution, according to Forman and Kedar (2004). The authors place that date in the 1948 Absentee Property Regulations, which the March 1950 statute replaced and carried forward.

Why did the UN value refugee land as at 29 November 1947?

The UN Conciliation Commission’s 1961 working paper states that the date “was chosen because it was the last date before the exodus of refugees when land values in Palestine were reasonably stable.” Every component of the valuation, rural, urban and Jerusalem, uses that same date.

How much land was under Jewish or state ownership in 1948?

About 2.8 million dunams, roughly 13.5% of the 20.6 million dunams inside Israel’s borders at the time, were under formal state or Jewish ownership, according to Forman and Kedar (2004). Of Mandate Palestine’s 26.3 million dunams as a whole, 1.5 million were Jewish-owned by 1948.

Did the UN’s 711,000 include Palestinians displaced inside Israel?

No. The Technical Committee’s 1949 estimate counts “refugees from Israel-controlled territory” (United Nations Conciliation Commission for Palestine, 1950), so it excludes people displaced within Israel. No figure for internal displacement can be derived from it.

Which body handled the abandoned-land valuation?

The Refugee Office of the UN Conciliation Commission for Palestine arrived at the evaluation in September 1951, and the Commission’s Secretariat reported it in the 1961 working paper. The paper covers the compensation question from 1948 to 1961 across six annexes.

How did the Development Authority fit into the transfer of land?

Forman and Kedar (2004) describe it as an intermediary body, legally defined as extragovernmental. It could buy absentee land only from the Custodian and sell it only to the state and the Jewish National Fund. Its holdings were consolidated into “Israel Lands” under the July 1960 legislation.

Sources

• Forman, Geremy, and Alexandre Kedar, 2004. From Arab Land to “Israel Lands”: The Legal Dispossession of the Palestinians Displaced by Israel in the Wake of 1948. Environment and Planning D: Society and Space 22(6), 809 to 830. DOI 10.1068/d402. Peer-reviewed.
• Israel Law Review, 2018. 1948 Refugees: Proceedings of an international workshop, Hebrew University of Jerusalem Faculty of Law, 14 to 15 December 2016. Israel Law Review 51(1), 47 to 110. Cambridge University Press. DOI 10.1017/S0021223717000280. Workshop proceedings (rapporteur’s summary), not peer-reviewed.
• Morris, Benny, 2004. The Birth of the Palestinian Refugee Problem Revisited. 2nd edition, Cambridge Middle East Studies 18. Cambridge University Press. ISBN 978-0521009676. Scholarly book.
• Pappé, Ilan, 2006. The Ethnic Cleansing of Palestine. Oneworld Publications. Scholarly book.
• United Nations Conciliation Commission for Palestine, 1950. General Progress Report and Supplementary Report of the United Nations Conciliation Commission for Palestine, covering 11 December 1949 to 23 October 1950. UN General Assembly Official Records, Fifth Session, Supplement No. 18, A/1367/Rev.1. https://www.un.org/unispal/document/auto-insert-180810/. Official record.
• United Nations Conciliation Commission for Palestine, 1961. Historical Survey of Efforts to Secure the Implementation of Paragraph 11 of General Assembly Resolution 194 (III): Question of compensation. A/AC.25/W/81/Rev.2. https://unispal.un.org/pdfs/3E61557F8DE6781A052565910073E819.pdf. Official record.