Missing millions: Java’s 1944–45 famine in Indonesia’s historiography

Abstract

This article asks why the victims of the 1944–45 famine in Indonesia’s main island of Java are largely missing from Indonesia’s public memory and historiography. It surveys relevant studies, to conclude that there is no consensus on the human toll of the famine. The article then traces the origins of an initial estimate of four million mentioned by Indonesia’s authorities to data on mortality and births uncovered in late 1945. It discusses the outcomes of a recent study that analysed these data to re-estimate excess deaths of, respectively, 0.7 and 1.2 million during 1944 and 1945. The difference with the initial estimate is that it also included unborn children and an unsubstantiated approximation of victims in 1946. The article analyses the likely reasons why the millions of victims of the famine went missing from Indonesia’s public memory and historiography during the 1950s and 1960s.

Introduction

In September 1951, Indonesian foreign minister Achmad Subardjo addressed representatives of 52 nations at the San Francisco conference to conclude a postwar international peace treaty with Japan. He stated: ‘The damages which Indonesia suffered during the Japanese occupation are twofold: first, the loss of life of approximately four million people; and second, material damages of billions of dollars.’¹ Indonesia’s Ministry of Information republished Subardjo’s speech in the Indonesian language that same month.² But newspapers in Indonesia did not report Subardjo’s statement on the human toll of the occupation years: it was apparently common knowledge and not newsworthy. However, if these four million casualties had all been victims of the 1944–45 famine in Indonesia’s main island of Java, the island would have lost about 8 per cent of its 1940 population.³ In that case, Java’s human toll would have exceeded that of the widely publicized 1943–44 Bengal famine, when 2.1 million or 3.5 per cent of Bengal’s 1941 population died of starvation.⁴

The main purpose of this article is to identify the origin of this estimate of four million victims and establish reasons why it is not part of the current historiography of Indonesia during the Japanese occupation years. The next section notes that the estimate disappeared from public discourse in Indonesia and that it took until the 1980s for historians to revisit the demographic consequences of the 1942–1945 Japanese occupation of Indonesia, but that historiography has yet to reach a consensus on the human toll. Consequently, the 1944–45 famine is barely mentioned as one of Indonesia’s and the world’s twentieth-century tragedies. The third section dispels the suggestion that Indonesian representatives exaggerated this toll in order to maximize the claim for war reparations from Japan. The article then traces the origins of the estimate back to late 1945, when unpublished vital statistics for 1943–1945 became available for analysis. It discusses the outcomes of a recent study that corrected and augmented these data in order to estimate the net loss of people in Java during 1942–1945 due to increasing mortality and falling births. The section that follows asks why the victims of the 1944–45 famine in Java disappeared from Indonesia’s public memory and its national historical narrative.

The human toll of the Japanese occupation of Indonesia in historiography

After Subardjo’s 1951 speech in San Francisco, the estimate of four million casualties during 1942–1945 disappeared from public discourse.⁵ The possible reasons will be discussed below. It took more than 30 years for historians to revisit the question of the human toll of the Japanese occupation of Indonesia.

In a multi-volume study of Indonesia during the 1940s, Dutch historian Louis de Jong in 1985 described the evidence of increasing food shortages and impoverishment in Java during 1944–1945. He referred to the published estimate of ‘a statistician’ of 2.45 million deaths during the Japanese occupation years.⁶ This ‘statistician’ was agricultural economist Egbert de Vries, a high-ranking official of the Department of Economic Affairs in Jakarta during the 1930s and professor of economics at the nascent University of Indonesia during 1941–1942 and 1946.⁷ After his release from a Japanese detention camp in Jakarta in September 1945, De Vries acquired annual data on births and deaths at residency-level in Java during 1943 and 1944, as well as monthly birth and mortality rates for Jakarta residency during 1943–1945. He combined these data to estimate a net population loss, comprising higher than normal mortality and lower than normal births, of 2.45 million during 1943–1945. He published these findings in May 1946.⁸

During the 1980s, historians Aiko Kurasawa and Shigeru Sato also studied Java’s society during the 1942–1945 Japanese occupation.⁹ Their publications confirmed the difficulties in Java with food supplies, particularly during 1944 and 1945. However, they did not estimate the human toll of the 1944–45 famine.

In several publications, Kurasawa noted that net population growth in Java had been negative during 1944.¹⁰ Her 1981 and 1983 articles made explicit reference to the 1943–1944 birth and mortality rates published by De Vries. But her 1988 PhD thesis ignored this source and referred instead to an item from the archives of the Ministry of Colonial Affairs in The Hague, which contained the same rates as De Vries had published. She commented that these data ‘look very “unnatural” and the writer is very sceptical about their reliability’.¹¹ In 1993 she added that they ‘were probably compiled by Dutch colonial authorities in 1950’, even though De Vries had clearly specified the origins of the data.¹²

Sato also referred to the De Vries estimates, but did not comment on the significant excess mortality in 1943 and 1944.¹³ Instead, Sato expressed doubts about the death toll that Subardjo had mentioned in 1951, noting that the Indonesian authorities had ‘overestimated’ the toll for the purpose of claiming war reparations from Japan.¹⁴

Both Kurasawa and Sato referred only to the increased mortality rates during 1943 and 1944 mentioned by De Vries. Both omitted to mention that De Vries had estimated the human toll of the occupation years as a minimum of 2.45 million people in Java: 120,000 in 1943, 813,000 in 1944, and 1.5 million in 1945.

These diverging views on the human toll during the Japanese occupation of Java have persisted. For example, in a 2005 book chapter, Sato discussed the 1944–45 famine, but explicitly denied the human toll estimated by De Vries, stating that it was based on ‘unreliable statistics’.¹⁵ However, in the same book, two other historians, Harry Poeze and Henk Hovinga, contradicted Sato with references to De Jong’s 1985 book, which cited the De Vries estimate.¹⁶

During the 2000s, an Indonesian-Dutch-Japanese research project, supported by the Japanese government through its ‘Peace, Friendship and Exchange Initiative’, did not reconcile these contrasting views. The project’s multi-authored encyclopedia of Indonesia during the Japanese occupation contained a chapter by Sato, which explicitly denied the scale of the 1944–45 famine in Java: ‘Indonesia did not experience a catastrophic famine like those in northern Vietnam or north-eastern India, in which millions of people starved to death.’¹⁷ However, in the same book, historian Adrian Vickers contradicts this with an indirect reference to the De Vries estimate.¹⁸ The encyclopedia contains no discussion of the human toll of the 1942–1945 Japanese occupation of Indonesia, particularly the 1944–45 famine, except for an extensive discussion of the very sorry plight of the romusha. These were the at least 2.6 million labourers from Java forced by Japanese authorities to work under often-abominable conditions, of whom at least 200,000 perished.¹⁹

The 1944–45 famine does not feature in other historical studies of Indonesia during the Japanese occupation. Historiography in Indonesia discussed the suffering of the romusha, but ignored the famine. For example, the last two editions of the official history of Indonesia, published in 1990 and 2008, do not mention it.²⁰ An alternative national history of Indonesia is also silent on the 1944–45 famine.²¹

Foreign historians of Indonesia have mentioned the 1944–45 famine in Java, but have not analysed it in detail. For example, Merle Ricklefs referred to it, but omitted the estimated toll.²² Adrian Vickers mentioned it with an indirect reference to the De Vries estimate.²³ To substantiate, Vickers quoted from ‘A mute’s soliloquy’, an autobiographical publication by author Pramoedya Ananta Toer, containing the author’s observations of the famished people he witnessed during his travels through Java in early 1945.²⁴ A book-length study by Ethan Mark of Java during 1942–1945 stated that ‘material conditions too went into a steep decline’ during 1944–1945, noting that ‘the sprawled, silent, half-naked bodies of men and women too hungry, sick or exhausted to move another step … began to dot Java’s city streets and country byways’. He placed the blame on the Japanese authorities ‘who did not seem to care’ and the ‘general state of acquiescence among much of Indonesia’s elite’ and stated that three million people died in Java and one million in the rest of the country, but does not discuss the basis of these estimates.²⁵ Hans Pols substantiated Mark’s brief note with graphic descriptions of the deteriorating health of the population in Java during 1944, but was also unspecific about the famine’s human toll.²⁶

The paucity of discussion of the human toll in Indonesia during the 1944–45 famine in Java contrasts with other accounts of human suffering in Asia during the Second World War due to famines. For example, the famines in Bengal during 1943–44 and in North Vietnam in 1945 are both subjects of ongoing research.²⁷ They are also an integral part of the retelling of the national history of India and Vietnam. This paucity also contrasts with ongoing discussions and historical research of other episodes of human tragedy in Indonesian history, particularly the 1918–19 influenza pandemic, the 1945–49 war of independence, and the 1965–66 killings of possibly half a million Indonesians.²⁸

This raises two questions. Why do historians hold such very different views on the estimates of the human toll in Java during the Japanese occupation published by De Vries in 1946? And why have historians of Indonesia been reluctant to note the death toll of the Japanese occupation if Subardjo—a renowned lawyer and public intellectual in the 1930s, an adviser to the Japanese military administration of Java during 1942–1945, a nationalist who served in Indonesia’s revolution against Dutch colonial rule during 1945–1949, and Minister of Foreign Affairs of the Republic of Indonesia—was confident to mention four million victims on a world stage in 1951?

Human losses and Indonesia’s claim for compensation payments

Sato opined that Subardjo had quoted an estimate of the death toll that had been ‘overestimated’ in order to maximize Indonesia’s claim for war reparations from Japan.²⁹ Sato based this view on the unpublished recollections of Sarimin Reksodiharjo who, during 1951–1952, had been a member of the Indonesian delegation to Japan to negotiate compensation.³⁰ Sarimin’s recollections and Sato’s interpretation of them are incorrect for two reasons.

First, Sarimin recalls that it was the ‘death and disappearance’ of four million romusha and ‘intellectuals’ (‘tjendikyawan’) and their estimated annual earnings over an average working life of 20 years that substantiated a compensation claim for the loss of life of US$10 billion. Nowhere does Sarimin’s text state that the Indonesian delegation purposely overestimated this claim. Sato justified his judgement that the toll had been ‘overestimated’ by arguing that the four million romusha and intellectuals who allegedly perished according to Sarimin would have comprised an incredible 25 per cent of the male labour force in Java. However, Sarimin seems to have incorrectly assumed that the four million deaths mentioned by Subardjo in 1951 had been romusha and ‘intellectuals’, but the next section will explain that the four million were part of the general population. This negates Sato’s argument that the Indonesian delegation ‘overestimated’ the human toll on purpose.

Secondly, Sarimin is mistaken in his recollection that the estimate of war damage and the implicit claim for compensation submitted by the Indonesian delegation during the 1951–1952 negotiations in Japan included a monetary value associated with human losses. The damage estimate and the implicit claim was indeed a very high US$17.5 billion, but this sum did not include a line item associated with the loss of human life.³¹ It could not have, because compensation for the loss of human life was explicitly excluded from bilateral negotiations about Japanese compensation payments following the San Francisco treaty.

Compensation for human losses was an issue of public discussion in Indonesia during the lead-up to the San Francisco treaty conference. After the American organizers had clarified in July 1951 that the conference would not discuss reparations payments, just the peace treaty, there was extensive discussion in Indonesia’s parliament and newspapers about the country’s participation in the conference.³² Several parliamentarians wanted Indonesia’s presence to be conditional on the arrangement of reparations payments, and the Indonesian government proposed various amendments to the treaty text. Only after Subardjo had met with Japanese delegation leader Shigeru Yoshida before the conference and had received assurances from the Japanese delegation that payments for ‘damage and suffering’ would be settled through later bilateral discussions, did the Indonesian government authorize him to sign the San Francisco treaty.

Following the signing, Indonesian delegation member Muhammad Yamin stated in a press conference in San Francisco that Indonesia’s claim for compensation would be around 25 billion Rupiah, that the claim would be based on the ‘Hart report’ and the work of a Republican interdepartmental committee, and that the compensation would be settled in bilateral discussions.³³ This ‘Hart report’ had been prepared during 1946–1948 by a committee appointed by the colonial government of Indonesia to estimate war damage and claim war reparations from Japan. Its chair was Henri M. J. Hart, former head of the Central Office of Statistics (Centraal Kantoor voor de Statistiek) in Jakarta from 1938–1942. In 1948, this committee estimated war damage until the Japanese surrender in August 1945 to have been f 24.5 billion (in 1938 prices).³⁴ The Hart Committee based its work on the guidelines of the Far Eastern Commission, an advisory committee to the Supreme Commander for the Allied Powers. Its report included material losses of the Indonesian population, but not an estimate of the monetary value of the loss of human life.

Nevertheless, in the lead-up to the bilateral meeting on compensation in Tokyo, the media in Indonesia expected that the delegation would submit a demand for compensation for the loss of life as part of its claim for compensation payments. They widely quoted an Agence France-Presse report which mentioned that an unnamed Japanese source involved in the negotiations had stated that the Indonesian delegation had expressed this expectation.³⁵ However, three days later, a United Press agency report from Tokyo mentioned that delegation leader Djuanda Kartawidjaja had explicitly denied that Indonesia would ask for compensation for the loss of human life during the Japanese occupation.³⁶

The origins of the initial estimate of four million victims

Several studies have confirmed that food supply decreased in Java during 1943–1945 and that the number of people suffering malnutrition and other famine-related consequences increased. For example, Kurasawa and Sato analysed the situation in Indramayu regency (kabupaten) and Lucas investigated the Pemalang and Pekalongan regencies.³⁷ Studies have offered various reasons for the worsening food situation. In several publications, Kurasawa and Sato outlined an amalgam of factors, including deteriorating rail transport facilities, corruption among Indonesian regional public servants leading to underreporting of rice production and sabotage of the efficacy of the official rice purchase and distribution system, and the unwillingness of Javanese farmers to follow Japanese orders about how to plant rice. Van der Eng and Huff emphasized the effects of the rigid system the Japanese authorities in Java used to purchase rice from farmers for distribution purposes. Increasing restrictions on the free markets for food products, combined with paying farmers for rice deliveries with increasingly worthless paper money, minimized any incentives for farm surplus production beyond the subsistence requirements of farm households.³⁸ Brennan et al. argued that the extended dry season in 1944 and delayed main rice harvest of the 1944–45 cropping year aggravated the food situation in early 1945.³⁹

A reason for these different explanations is the paucity of direct information on the extent and causes of the 1944–45 famine in Java. Japanese censorship prevented the media in Indonesia from reporting anything related to this famine.⁴⁰ Nevertheless, inklings of drastic food shortages in Java reached the rest of the world during 1944 and early 1945. Representatives of the colonial government of Indonesia in exile, located in Australia, interviewed people who managed to flee Indonesia and arrived in West Papua or northern Australia. They also interviewed romusha from Java discovered by Allied forces in the course of April–May 1944 along the north coast of West Papua (Hollandia/Jayapura, Biak, Manokwari) and April–May 1945 in Maluku (Morotai) and North Kalimantan (Tarakan). These testimonies informed secret reporting by colonial government officials from Melbourne, particularly from March 1945.⁴¹ They also informed a summary publication which concluded ‘the food situation is desperate’ in Java.⁴² Several newspapers in Australia repeated this.⁴³

After the Japanese surrender on 15 August 1945, it took until 4 October for representatives of the colonial government to start returning to the urban enclaves of Jakarta, Surabaya, Semarang, and Bandung. Details of the very dire food situation in rural Java then started to trickle in. One of the objectives of the returning colonial government was to implement an extensive relief plan in Indonesia.⁴⁴ It involved the distribution of large stockpiles of relief goods accumulated by the colonial government in exile, including food. However, due to the political situation, its ships only reached parts of eastern Indonesia in late 1945 and the urban enclaves in Java, not rural Java.

Colonial government representatives predicted that the food situation would get worse in the lead-up to the main harvest in April–May 1946. For that reason, the director of the Department of Economic Affairs, Jacob E. van Hoogstraten, in a press conference on 19 October 1945, explained the consequences of strikes in Jakarta port against the unloading of Dutch ships bringing in relief supplies, and strikes by waterside workers in Australian ports against the loading of relief supplies acquired and stockpiled by the colonial government in exile in Australia for shipment to Indonesia. Van Hoogstraten was the first to mention that four million people in Java had died from starvation and diseases during the Japanese occupation.⁴⁵

In a public speech in Amsterdam on 4 January 1946, broadcast in The Netherlands and to Indonesia, the head of the colonial government, Lieutenant Governor General Hubertus J. van Mook, mentioned that two to four million people in Indonesia had become victims of ‘murder, hunger, deportation and epidemics’ in Java during the Japanese occupation.⁴⁶ Newspapers in Indonesia also reported on this aspect of Van Mook’s speech.⁴⁷

The only substantiation of these estimates of the death toll in Indonesia were the data that De Vries would publish in May 1946. His main source were the numbers of births and deaths by residency in Java during 1943 and 1944, which Indonesian nationalist Prawoto Soemodilogo would publish in his journal Ma’moer on 10 January 1946.⁴⁸

It is likely that sometime in late 1945 Prawoto had shared these data with De Vries. Both must have known each other during the 1930s. During 1931–1942, Prawoto had been a member of the People’s Council (Volksraad), the parliament of colonial Indonesia, in Jakarta. He had participated in several economic committees of the council during the 1930s. During 1934–1941, De Vries had worked in the Department of Economic Affairs in Jakarta, which was occasionally consulted by the committees of the People’s Council. Both attended the same Council meetings.⁴⁹ During the 1930s and early 1940s, Prawoto and De Vries both published on aspects of Java’s food economy and clearly shared that interest.

After August 1945, Prawoto did not become part of the government of the Republic of Indonesia. Instead, in December 1945, he started publishing his journal Ma’moer in Jakarta, which was very critical of aspects of the economic policies of the Republican government.⁵⁰ He may have renewed contact with De Vries in Jakarta after the latter had been released from Japanese detention in Jakarta, probably in September 1945, and before De Vries resumed his position as professor of economics at the nascent University of Indonesia in January 1946.

Prawoto was in a position to acquire the 1943 and 1944 vital statistics, because he worked for the Japanese military government. In 1944 he had been adviser to the Department of Economic Affairs (Sangyobu, 産経部) of the Japanese military government in Java. He investigated the April–August 1944 rebellions of farmers in Indramayu against the requisitioning of rice by Japanese authorities.⁵¹ As departmental adviser, he took part in the 1944–1945 Advisory Committee (Sanyo Kaigi, 参与会議), consisting of all Indonesian advisers to government departments, to inform the Japanese military government on several issues, including the system for acquiring rice for distribution purposes. During December 1944, he surveyed the food situation in Jakarta, Cirebon, and Pekalongan as part of a subcommittee that analysed the deteriorating food situation in Java. This sub-committee collected birth and mortality statistics in Central Java that indicated that mortality rates had started to exceed birth rates in late 1943.⁵²

In February 1945, the Japanese military government appointed Prawoto to the Central Advisory Council (Chuo Sangi-in, 中央参議院) in Jakarta.⁵³ Given his understanding of the dire food situation in rural Java at that time and his general interest in the rural economy, Prawoto may have obtained the 1943–1944 data on births and deaths in mid-1945 while working at the council.⁵⁴ He would have had relatively easy access to these data, because one of the tasks of the statistical section in the former colonial Public Health Service (Dienst voor de Volksgezondheid) in the Department of Home Affairs had been to receive health reports with the regional birth and mortality statistics, before aggregating them into annual averages by residency. The health service continued its operations during 1942–1945 as the Eiseikyoku (衛生局) and after August 1945 as the Djawatan Kesehatan.⁵⁵ The council and the health service were located in each other’s vicinity in Jakarta.

It is therefore very likely that Prawoto was the main source of the vital statistics that De Vries used in September or October 1945 to estimate the net loss of people in Java during the Japanese occupation. He may have done this just before Van Hoogstraten held his press conference on 19 October 1945, and well before January 1946 when Prawoto published the 1943–1944 vital data and Van Mook mentioned two to four million deaths.

When De Vries published his estimate of the 1943–1945 net population loss of 2.45 million in May 1946, he added ‘It is to be feared that this unfavourable development is still having its aftermath in 1946, and that the total direct and indirect war losses in Java amount to from 3 to 4 million souls.’⁵⁶ Consequently, the estimated four million victims included an unsubstantiated approximation of excess mortality in Java during 1946.

This number of four million victims reappeared several times in later years. In January 1947, a United Nations report on the war damage in Indonesia mentioned: ‘starvation, neglect of health services, requisitioning of medical stocks, the internment of physicians, lack of transport, and general deterioration of economic and social conditions resulted in a net loss of 4 to 5 million people who would have been alive if the natural growth of the population had not been interrupted’, adding: ‘of the half to one million Indonesians and Chinese used as forced labourers by the Japanese, about 300,000 died’.⁵⁷ A February 1948 report by the Dutch Minister for Overseas Territories to the Dutch parliament left no doubt that the 1946 De Vries article was the basis for the estimated loss of people during 1943–1946.⁵⁸ This report and the estimated toll of four million were widely reported in newspapers in The Netherlands and in Indonesia.⁵⁹ Consequently, the estimate of four million victims was public knowledge in Indonesia at the time of Subardjo’s San Francisco speech in September 1951.

A conjoint article scrutinized the quality of the available birth and mortality data during the 1930s–1950s and concluded that both were largely underestimated, to the extent that infant mortality was under-recorded.⁶⁰ It augmented and corrected the data for this under-recording on the basis of the work of Indonesian demographer Widjojo Nitisastro.⁶¹ Figure 1 shows the results, which indicate elevated mortality rates during 1944–1951 and decreased birth rates during 1944–1948 compared to the late 1930s and early 1950s, and a net negative population growth during 1944–1948, particularly in 1945. These estimated rates allowed a re-calculation of total excess deaths and missing births, which amount to a net population loss of 3.4 million during 1942–1945.⁶² This estimate consists of 1.9 million excess deaths (0.7 million during 1944 and 1.2 million during 1945), and a residual of 1.5 million missing births.⁶³

Note: Refers to only the Indonesian population in Java.

Figure 1. Average birth and mortality rates in Java, 1936–1955 (‰). Source: Pierre van der Eng, ‘Mortality from the 1944–1945 famine in Java, Indonesia’, Centre for Economic History Discussion Paper No. 2024-01 (Canberra: ANU College of Business and Economics, 2024).

The 1944–45 famine and Indonesia’s public memory and national narrative

Any answer to the question of why the estimate of four million victims of Japan’s 1942–1945 occupation of Indonesia disappeared from Indonesia’s public discourse and national memory after 1951 has to be hypothetical. There is as yet no documentation that substantiates a definitive answer.

As mentioned above, knowledge of the disastrous famine in Java reached the rest of the world and was published in March 1945. But within Java, various impressions of disastrous situations had started to accumulate earlier, during 1944 and 1945. Due to Japanese censorship, these were not publicly shared via newspapers, but impressions were later harvested by historians (see above). What did those in public office know during 1944 and 1945 about the unfolding famine and what did they do to alleviate the situation?

Before 1942, and also during the 1950s and early 1960s, regional health service officials used hospital admissions and the number of recorded hunger oedema sufferers, as well as weekly regional mortality rates, as indicators to alert central authorities to regional health crises, including famines, and call for alleviating interventions. This may still have happened during 1942–1945. However, the censored newspapers did not publish such details, and historians have so far not found archival records to allow an assessment of what particularly the Japanese military government knew in relation to famine threats.

The details of regional human consequences of the 1944–45 famine that historians have been able to uncover and reconstruct, suggest that at the time knowledge of the unfolding disaster was patchy and limited, but would have been clear to anyone who was in a position to interpret this limited official information. Following the dismissal and detention of Dutch public officials from the public service in 1942, the main keepers of this information were Indonesian officials employed in regional offices or in central government departments who carried out existing public functions or who took new positions created by the Japanese military government.

Local officials would have generated the data that would have indicated the unfolding disaster, particularly the weekly round-ups of hospital reports and vital statistics, as well as the monthly reports on the area planted with farm crops and harvested. The group of local officials overlapped with the smaller second group, which included leading Indonesian senior public officials, academics, as well as medical and legal professionals, many of them ardent nationalists. From 1943, many took part in the Chuo Sangi-in advisory council and its regional equivalents. Although the Japanese military government limited the authority of the advisers and their access to administrative information, the work of these organizations indicates what information, however patchy, was shared between Indonesian and Japanese public officials. The advisory organizations would have had opportunities, however limited, to alert the Japanese authorities to such unfolding issues.

The task of the advisory organizations was to respond to specific questions from the Japanese authorities at the central or regional levels, rather than provide unsolicited advice. For one of those organizations we have further details: the Sanyo Kaigi advisory committee which comprised the senior Indonesian advisers to each department of the military government. It had been established on 11 November 1944 and was chaired by Sukarno, who later became president. Unlike the council, its deliberations and advice were secret, and its discussions must have been more frank. The purpose of the committee was to provide advice on specific measures that the military government intended to implement.

One of the measures was the system the government used to purchase paddy for distribution purposes. Purchases in 1943 and 1944 had been below expectations, which limited the rice available for distribution. This issue had been discussed in November 1944 in the Chuo Sangi-in advisory council, which agreed that the paddy purchase system had to be improved. On 3 December 1944 the government asked the Sanyo Kaigi for more specific advice, before the reorganized system would be implemented during the main rice harvest in 1945. The instructions from the head of the military government, Moichiro Yamamoto, to the committee included that its proposals would have to be guided by this consideration: ‘Can I bear responsibility for this if I would have to carry out this proposal myself?’⁶⁴ Implicitly, he expected the committee members to share responsibility for any measures that the government would take in relation to the recommendations.

When the committee met to discuss this issue on 16 December 1944, its members must have known that the food situation in Java had deteriorated in the course of that year. First, because several committee members were also members of the Chuo Sangi-in advisory council, including Sukarno who chaired both fora. The council had discussed the issue in November 1944 to conclude that ‘the people’s physical health and strength would decline if the food supply were not optimally administered’.⁶⁵ Secondly, senior public official Ahmad Subardjo had travelled throughout Java in early 1944 and his July 1944 report had analysed the dire food situation, mentioning increasing food shortages and malnutrition in Purwokerto, Bojonegoro, and Semarang.⁶⁶ Thirdly, in June 1944, council committee member and medical doctor Boentaran Martoatmodjo had published an analysis of the increasing mortality rates in Semarang residency during 1943 and 1944 in the medical journal Berita Ketabiban, finding increasing malnutrition to the extent that in 1943 mortality rates had started to exceed birth rates.⁶⁷

The committee decided to send six of its members, including Prawoto and Boentaran, to different regions in Java to investigate the food situation and report back for the committee’s next meeting on 8 January 1945, which discussed these regional reports and provided answers to the specific questions from the military government.⁶⁸ The regional reports underlined the growing urgency of food shortages in Java. Most concerning was Boentaran’s report, which updated his 1944 article with a frank discussion of the food and health situation in Semarang residency and reported similar mortality and birth rates in Kedu and Pati residencies during 1944. The committee summarized the reports thus: ‘the people’s physical stamina and ability to resist disease has sharply declined. In general the death rate now exceeds the birth rate.’⁶⁹ Apart from Boentaran, only Prawoto seems to have realized the gravity of the health situation and the disaster that was unfolding in rural Java. Prawoto stated: ‘I was dumbfounded to hear Dr Boentaran’s report on the rising death rate and the declining birth rate.’⁷⁰

Consequently, in January 1945, committee members knew that Java’s population was already badly afflicted by malnutrition and increasing mortality. The purpose of the meeting was not to dwell on the worsening health situation, but to provide advice on the paddy purchase system. Nevertheless, these issues were related, because the result of shortcomings in the purchase system meant that insufficient rice was available for distribution to alleviate regional food shortages.

Committee members had very different views on the reasons for shortcomings of the paddy purchase system. The discussion was not informed by any data on the regional farm areas under food crops and food production, even though such data would have been available.⁷¹ The discussion was mostly based on casual observations of paddy and rice hoarding, malfeasances by Chinese rice mill owners, and the existence of a black market with high rice prices. Implicit was the notion that actual paddy production was sufficient in Java, but that farmers and speculators were hoarding paddy, causing regional rice shortages and obstructing the operations of the paddy purchase and rice distribution systems.

None of the participants questioned the legitimacy of the paddy acquisition system, but views differed widely on how it should be implemented. Economist Muhammad Hatta argued against using regional paddy quotas and proposed that the system only acquire all the paddy that was surplus to subsistence requirements in regions in order to minimize the impact on rural food supplies. In stark contrast, Oto Iskandardinata believed that the system had to acquire Java’s entire paddy crop, before distributing the rice fairly, including to farmers who had produced it in the first place. And Abikusno Tjokrosujoso went one step further, arguing at length that all production in Java had to be centrally controlled by a directorate (‘direktorium’), an idea that Sukarno summarized as a dictatorship (‘diktator’).⁷²

Although the meeting also proposed an increase in purchase prices, it did not query the simple fact that the system’s low real purchase prices had reduced the incentives for farmers to produce a surplus of paddy, which had made it difficult to meet the 1943 and 1944 regional purchase quota.⁷³ The tenet of the meeting was to identify ways to mitigate any resentment of farmers against the purchase system, of which there had been several instances (as noted above) and against the Japanese military government. Inter alia, the committee ignored that Java’s farmers had agency in the whole issue, agreeing instead that farmers had to be obedient to the government and surrender their paddy to the whole population for compassionate and patriotic reasons.

In the end, the committee voted down the Abikusno and Hatta proposals and remained equally divided over Oto’s unrealistic proposal.⁷⁴ It accepted recommendations to nationalize all rice mills and increase the paddy purchase prices. It also decided to submit the regional reports, the minutes of the meeting, and its recommendations to the military government. Implicitly, the meeting opted not to use this opportunity to highlight its findings of malnutrition and excess mortality as indicators of the unfolding famine in Java in 1945 and to stress the urgency of practicable solutions for relieving the food situation.

On the other hand, except for any casual observations of the worsening food situation, especially in urban areas to which famished people fled, it would still have been difficult in January 1945 to extrapolate the adverse findings from Central Java to the rest of the island.⁷⁵ The vital statistics for all regencies in Java for the whole of 1944 still had to reach the residency offices, before they would arrive in Jakarta by mid-1945. Regional monthly reports on planted areas with food crops during the October–December 1944 start of the 1944–45 main cropping season would still have been in the processing phase, before they would be sent to the central statistical office in Jakarta. Consequently, based on what was known in early January 1945, committee members would have had grounds, if required, for later plausible denials of any co-culpability for the unfolding famine.

As things were, the committee or council members did not have to deny knowledge of the 1944–45 famine, because it took until January 1946 for Prawoto to publish the 1943–1944 vital statistics and until May 1946 for De Vries to publish his estimates of the net loss of 2.45 million people during 1942–1945. By then, Indonesia was engulfed in the struggle for full independence. On both sides of the conflict, authorities were dealing with more urgent matters than revisiting the 1944–45 famine.

Nevertheless, there was public discussion in Indonesia about the culpability of particularly President Sukarno and other leading Indonesian nationalists who, in their opinion, had collaborated with the Japanese military government in the human suffering of 1942–1945. For example, since his return to Indonesia in 1945 Indonesian communist leader Tan Malaka was openly dismissive of Sukarno’s close cooperation with the Japanese.⁷⁶ In September 1948, in a public broadcast, Munawar Musso, the leader of a communist uprising in Madiun, held Sukarno accountable for sending hundreds of thousands of romusha to their deaths and—by his account—widowing the wives of two million romusha.⁷⁷ In 1951, leading Jakarta newspaper Indonesia Raya (2 February 1951) published a letter to the editor in the column ‘People’s thoughts’ (Fikiran orang banjak). Its author A. Kamil held Sukarno culpable for the deaths of the romusha. After Sukarno lodged a formal complaint with Indonesia’s attorney general for ‘hurt feelings’ (perasaan terganggu), the paper’s editor refused to apologize for the publication of the letter, arguing that its intent was not to insult the president and was based on historical facts.⁷⁸

Sukarno would in 1965 take responsibility for the deaths and suffering of the romusha.⁷⁹ However, he justified his role as being part of a greater good—the process that eventually led him to declare Indonesia’s independence in August 1945. A reason for Sukarno taking responsibility was that the plight of the romusha could not be denied, because he had played—due to Japanese propaganda—an active and very visible role in their recruitment.⁸⁰ In addition, during the 1950s, surviving romusha self-organized in various associations for mutual support and in order to keep the public memory of their dead compatriots alive. And the government of Indonesia acknowledged the suffering of the romusha through public memorials, such as in Pekanbaru (Sumatra) and in Bayah (West Java) in relation to the Muaro-Pekanbaru and Saketi-Bayah railways.

This explains that the plight of the romusha continued to be part of Indonesia’s national historical narrative; it has been repeatedly publicly articulated. By contrast, the fate of the starving millions in Java and fact that Indonesian leaders advising the Japanese military government knew about this was not publicly articulated. Consequently, while the human toll of four million was public knowledge in 1951, as was concluded above, the culpability of leading Indonesian nationalists in terms of condoning the paddy acquisition system that aggravated the food situation and the 1944–45 famine was not publicly known. The deliberations and recommendations of the Sanyo Kaigi took place in secret and were not trumpeted in Japanese propaganda. In addition, there is no evidence that people who survived the 1944–45 famine self-organized to articulate their plight.

In the course of the 1950s, opportunities to hold particularly Sukarno accountable for the plight of the romusha and other human tragedies during 1942–1945 decreased. One reason is that it was no longer necessary to refer to the plight of the romusha and the four million victims in order to put pressure on Japanese counterparts during discussions about compensation, after it had become clear in December 1951 that compensation for human losses would not be part of Indonesia’s implicit claim.

A further reason is that Indonesia experienced an end to its short phase of rough and tumble parliamentary democracy and a decline in press freedom after Sukarno declared martial law, abolished parliament, appointed himself executive president, and increased censorship on newspapers in March 1957, then banned the publication of various newspapers (including Indonesia Raya) in October 1957, and imposed ‘guided democracy’ in 1959.⁸¹ This was accompanied by the increasing marginalization and silencing of Sukarno’s political adversaries, and an invigorated propaganda that shielded Sukarno’s reputation as the father of Indonesia’s revolution and independence in the 1940s. Consequently, after 1957 Sukarno no longer had to lodge defamation claims in relation to the romusha or other human tragedies during 1942–1945; Indonesia’s censors would have pre-empted this.

In addition, during the 1950s, Indonesia’s economy experienced increasing difficulties. These were essentially fuelled by runaway budget deficits that caused accelerating inflation, and an exchange rate regime that slowed Indonesia’s export earnings and increased its trade deficit. In that light, the Indonesian government became increasingly keen to reach a settlement in the negotiations with Japan about compensation payments. Apart from an inflow of payments, the government expected the agreement to be the basis for the normalization of trade relations, as well as new investment by Japanese companies in Indonesia at a time when foreign investment from other parts of the world was flagging. In these contexts, there was no longer any incentive for anyone in Indonesia to scrutinize the deplorable nuances of the 1942–1945 years in public, let alone identify culpability.

The bilateral Indonesia-Japan negotiations about compensation payments first took the form of the Djuanda mission to Tokyo during December 1951 and January 1952, which failed to reach a final agreement. Further rounds of negotiations did not conclude until January 1958, when both parties signed a bilateral peace treaty. One of the reasons for the slow progress was Indonesia’s high implicit initial claim of US$17.5 billion. In addition, Japan intended to provide compensation payments in kind, essentially in the form of construction and development projects, rather than cash. Between 1952 and 1958, several negotiation rounds whittled down Indonesia’s compensation claim to US$8 billion in 1953, then to US$1.2 billion in 1956.⁸² In January 1958 both countries signed an agreement specifying US$0.8 billion in the form of Japan financing various development projects valued at up to US$223 million, annulment of Indonesia’s trade debt to Japan of US$117 million, and US$400 million of new Japanese loans to the Indonesian government.⁸³

When the Japanese war compensation started to flow, it did not directly benefit the families of the 1942–1945 victims. It consisted of various projects built by Japanese construction companies, including several that upon completion in the 1960s benefitted Indonesia’s elite, such as the Sarinah department store in Jakarta and four luxury hotels (Samudra Beach hotel in Pelabuhanratu, Ambarrukmo hotel in Yogyakarta, Hotel Indonesia in Jakarta, and Bali Beach Hotel in Sanur). Indonesia-Japan business relations improved after 1958. Indonesian state-owned companies increased exports of oil, bauxite, and other minerals to Japan; Japanese firms increased their exports to Indonesia; other Japanese companies started to invest in mining and manufacturing projects in Indonesia; and Japan’s foreign aid to Indonesia continued to grow.⁸⁴

Whether as a result of increased censorship since 1957 preventing any press reports that could stain Sukarno’s reputation or offend any perceived Japanese sensitivities at a time of Japan’s growing economic importance for Indonesia, or whether there simply was a multitude of more pressing current issues that dominated public discussion, the human toll of four million that had been public knowledge in 1951 faded from public memory in Indonesia and was no longer mentioned. Indonesia’s demographers, particularly Nitisastro, only implicitly researched the human toll in their reconstructions of population trends, without estimating excess mortality during the 1940s. And, as indicated above, Indonesia’s historians ignored it altogether. Without any reference to the extent of the human toll and without probing research into the key factors that worsened the food situation, it was (possibly by stealth) assumed that only the Japanese military government was accountable for the human suffering related to Java’s 1944–1945 food shortages.

Conclusion

This article has shed new light on a controversial issue in Indonesia’s historiography, as the second section explained. The controversy is related to the fact that the available demographic data have not been analysed in sufficient depth in order to reconcile the very different views on the demographic consequences of the 1944–45 famine in Indonesia’s main island of Java. The fourth section established that the estimate of four million victims, which was common knowledge in Indonesia around 1950, was based on the system whereby mortality and birth statistics were collected in Java. A closer look at these data identified a net population loss of 3.4 million during 1942–1945, consisting of 1.9 million excess deaths (0.7 million during 1944 and 1.2 million during 1945), and a residual of 1.5 million missing births.

This net loss estimate may seem similar to the four million victims mentioned by Subardjo in 1951, but it is different, because the latter includes the net loss of people in the aftermath of the Japanese occupation period in 1946. The estimate of excess deaths is higher than De Vries estimated in 1946, but differs because De Vries did not account for an underestimation of birth and mortality rates and included missing births. The previous section hypothesized the reasons why the common knowledge in the early 1950s of four million victims disappeared from Indonesia’s national memory. It elaborated that Indonesia’s guardians of public memory may have refrained from reiterating the human toll while the Indonesia-Japan compensation discussions were ongoing, and that from 1957 onwards Indonesia’s censors shielded Sukarno’s reputation in terms of his role during the Japanese occupation years.

These millions missing from Indonesia’s historiography matter to Indonesia’s society today for at least two reasons. First, they set the record straight: the Japanese occupation was not only a political and social watershed in Indonesian history, it was also an episode of catastrophic human suffering well beyond the plight of the romusha. The distressing fate of a very large number of ordinary Indonesians during the 1944–45 famine in Java is not widely acknowledged in Indonesia. Past major famines have become part of the national historical narrative in other countries, such as Ireland, Ukraine, India, and Vietnam, where the fate of the victims is widely acknowledged and commemorated. These victims are now part of the collective memory that contributes to how people in these countries define their national identity. By contrast, Indonesia’s collective memory currently lacks this.

Secondly, the demographic consequences of the famine era must have had lasting physical consequences which may linger until today and require further research. Studies of famines in other countries have identified the personal tragedy of families whose children died or who were unable to raise the children that were not born during these years. Such studies also identified the adverse physical and mental health effects on famine survivors, as well as the transgenerational and epigenetic effects due to in-utero genetic changes on the second and third generations of survivors of other famines. Examples of such studies relate to the famines in Ireland during 1845–52, The Netherlands in 1944–45, North Vietnam in 1945, and China during 1959–61.⁸⁵ No comparable studies exist in Indonesia, possibly because the millions of victims of the 1944–45 famine are still missing from Indonesia’s historiography.