.png)
Sino-Vietnamese War Casualties: The Count That Only Ran One Way (Cold Files 024)
On 17 February 1979 China invaded Vietnam. Sixteen days later Beijing announced it was withdrawing and declared the objective achieved. It said it had put roughly 57,000 Vietnamese out of action. Hanoi said it had put 62,500 Chinese out of action. Neither government said, then or since, how many of its own soldiers it had buried.
Forty-seven years later the figures a reader will find have a traceable ancestry, and almost none of it leads back to a count. One descends from a single book written in 1987 by a scholar working from press reports. One descends from internal Chinese army ledgers that one historian read, and then revised upward by 43 percent. The Vietnamese figures have no lineage.
The clearest evidence for how that happened is not propaganda out of either capital. It is the Chinese army's own classified paperwork.
Part of the Cold Files series.
The war that each side measured in the other side's dead
The scale has never been agreed either. Estimates of the invading force run from about 200,000, the figure the historian Xiaobing Li gives, through the twenty-seven divisions and 220,000 troops counted by Miles Maochun Yu of the Hoover Institution, to the "more than 300,000" Xiaoming Zhang gives for the committed armies. Vietnam's news agency says China mobilised 600,000.
The stated causes were two. Through 1978 some 160,000 ethnic Chinese, the Hoa, crossed overland out of Vietnam into China, and by mid-1979 China had taken in around 250,000 people from Vietnam, figures the journalist Barry Wain set out in Foreign Affairs that September. Then in December 1978 Vietnam invaded Cambodia and installed a government in Phnom Penh on 8 January 1979.
Three weeks later Deng Xiaoping was in the Oval Office. The American record of the meeting of 30 January 1979, printed in the Foreign Relations of the United States series, has him saying that "China must still teach Vietnam a lesson" and that "the action will be very limited". The day before he had told Carter, Mondale, Vance and Brzezinski that the plan was "a limited short time action, to give them a lesson", and that Chinese troops "will quickly withdraw".
What the planners had not counted was the militia. Zhang records 40,000 to 50,000 of them around Cao Bang alone, a force Chinese planning had underestimated badly enough that the local ratio collapsed from roughly eight to one down to two to one. Vietnam's best divisions were in Cambodia.
Beijing's public accounting of its own losses, in the whole of 1979, was a single number: about 20,000 killed or wounded, carried by Agence France-Presse out of Hong Kong on 2 May and generally traced to General Wu Xiuquan. One aggregate, no breakdown, against a precise claim of 57,000 enemy.
What the secret reports counted
In 2009 the historian Edward O'Dowd published translations of five Chinese army after-action reports on the 1979 campaign. They were written by officers who had served in it, compiled by the Military Academy's Training Department between 16 March and May 1979, classified jimi, and not authorised for further distribution. They were not written for foreigners. They were not written for the public. They were written for the instruction of the army that had just fought.
O'Dowd's finding about them is the hinge of this file. Each report's introduction, he writes, "lists the number of enemy troops killed, wounded, and captured" along with the enemy equipment taken, with results "stated with great precision". The same reports "do not routinely list the numbers of casualties incurred to accomplish their missions", and "do not list victims of friendly fire except in a few cases".
The precision is real and it is strange to read. The 163rd Division reported 5,293 Vietnamese killed and 612 of its own dead, and noted that the enemy figure excluded those killed inside bunkers and caves. The caveat runs one way only: the unit qualified the number it had an incentive to inflate and left its own standing bare.
Own losses were not absent from the paperwork everywhere. Zhang cites a 55th Army after-battle report giving 1,271 killed and 3,779 wounded at Dong Dang and Lang Son, and the military regions kept medical ledgers. The numbers existed at unit level. They were simply not what a report was for.
Deng himself closed the campaign on that note. Addressing party, government and military leaders at the Great Hall of the People on 16 March 1979, he declared victory, called the army's losses small, and remarked that its count of Vietnamese wounded might not be accurate. The man who had ordered the war doubted the enemy figure that vindicated it.
Where the Western number came from
For three decades the standard English-language figures were 25,000 Chinese dead and 37,000 wounded. They come from pages 113 and 114 of King C. Chen's China's War with Vietnam, 1979, published by the Hoover Institution Press in 1987. Chen had no access to Chinese internal records. He was reasoning from Western press reporting and intelligence-derived material. Zhang notes drily that Chen's casualty statistics "have been widely cited", and he is right: a version of that pair still sits in the general reference works a reader reaches first.
Then the archives half-opened, and the number moved. In The China Quarterly in December 2005 Zhang wrote that recent Chinese sources "categorize the PLA's losses as 6,900 dead and some 15,000 injured". In Deng Xiaoping's Long War, ten years later, he gave 7,915 killed and 23,298 wounded, a total of 31,213. That is a rise of 43 percent in the total, and the book does not explain it.
The 2015 figure is not an estimate. It is a sum. Guangxi accounts for 5,103 dead and 15,412 injured, Yunnan for 2,812 and 7,886. Add them and you get Zhang's totals exactly. The authoritative Western figure for Chinese dead is two military-region ledgers added together.
Which raises the hardest objection to this whole file, and it deserves stating plainly rather than burying. If those ledgers are a genuine itemised count, assembled at the time by the army that took the losses and reproduced in a General Political Department anthology, then the dead were counted. What is missing is not a count but a press conference, and that is a much smaller story.
Three things keep it from being that small. Zhang's own two totals contradict each other by 43 percent with no stated reason. Fravel notes that one of Zhang's sources was a Chinese military-enthusiast website that has since gone offline, so the most-cited casualty figure in the field now rests partly on material nobody can re-examine. And the Vietnamese side of the ledger is genuinely empty, so even a settled Chinese number leaves the war's dead unfixed.
Vietnam's own published figure arrived forty years late. In February 2019 the state news agency, quoting Dr Tran Huu Huy of the Vietnam Association of Historical Sciences, said Vietnam had suffered more than 30,000 military killed and injured and tens of thousands of civilian deaths. No breakdown, no basis, no method.
The decade nobody counted
The war did not stop in March 1979. China kept twelve divisions on the border and the raids continued for five years. Then in April 1984 it came back in force. O'Dowd counts more than 60,000 shells fired into sixteen districts across four Vietnamese provinces that month, before the ground assault went in at dawn on 28 April and took the heights above Thanh Thuy, in Vi Xuyen district of what was then Ha Tuyen province.
Vietnam counter-attacked on 12 July 1984, in an operation its veterans still mark by that date. More than 600 men of Division 356 were killed in a single day. The soldiers' names for the ground have outlasted the official ones: lo voi the ky, the century's lime kiln, because the shellfire burned the limestone white, and coi xay thit nguoi, the human meat grinder. Chinese records logged 1,080 Vietnamese dead that day and then revised the figure to 3,000 after intercepting Vietnamese signals.
For the Chinese army the hills became a syllabus. Xiaobing Li records armies from seven military regions rotating through the Laoshan sector for combat experience, a posting known as "touching the tiger's butt". O'Dowd counts at least fourteen armies passing through between 1984 and 1989. Zhang's figure, as cited by M. Taylor Fravel, is more than 180,000 Chinese soldiers over a nine-year period.
For that decade China has produced no count of its own dead. The figure everyone now prints for the Vietnamese side, more than 4,000 killed and more than 9,000 wounded, came out of the founding meeting of a veterans' liaison committee at the Ministry of National Defence in Hanoi on 14 July 2016. Major General Nguyen Duc Huy, the front's former chief of staff, gave the 4,000; the Ha Giang party secretary added the 9,000. And state media have printed it ever since. It is a veterans' number that the state adopted. It is not a number the state produced.
The cemetery constrains it. The historian Martin Grossheim records 1,797 soldiers buried at Vi Xuyen, about 1,500 of them from this war, and around 330 sets of remains there still unidentified. State media put 1,000 to over 2,000 still up there, in ground too heavily mined to work.
The cemeteries that do not name the war
What replaced the counting was a liturgy with the subject removed.
At Malipo in Yunnan, a cemetery begun in 1979, every headstone carries a portrait and a biography. It is a provincial patriotic-education base. Its official description says the dead there sacrificed to defend the frontier, and does not name the war or the enemy. Chinese school textbooks barely mention the war. The Chinese foreign ministry's own overview of relations with Vietnam omits it, while covering other border conflicts.
In China the memory came back instead as a benefits claim. More than 3,000 veterans reached the Friendship Pass at Pingxiang on the anniversary in February 2014 and had their banners taken within minutes; an organiser, Cai Jian, said the authorities "gave us a time limit, so there was little we could do within that time". Several hundred had gathered in Guangzhou in July 2012 over a monthly allowance of 510 yuan.
Vietnam's silence was negotiated. Grossheim records that at the Chengdu meeting of September 1990 Vietnamese leaders reportedly undertook to play down commemoration of the border war in the media. In 2014 the Historical Sciences Association was not permitted to hold a conference on the thirty-fifth anniversary. At least ten people were detained at unofficial commemorations on 17 February 2017.
Then on 5 February 2025 the party's general secretary, To Lam, offered incense at the Vi Xuyen national martyrs' cemetery, with the defence and public security ministers present. The government's own report of the visit names neither China nor the war. It says only that the dead there fell defending the northern border.
The state is still bringing them down. Under a 500-day national campaign, teams at Vi Xuyen had recovered 146 sets of remains, 98 of them from eleven collective graves, by late September, having searched some 2,750 hectares and moved over 27,000 cubic metres of soil and rock.
Forty-seven years on, two armies have between them produced one aggregate, two provincial ledgers, a veterans' committee announcement and a 1987 book. The convention that published only the other side's dead has outlived every reason either state had for it.
Sources consulted
O'Dowd, PLA Documents, Chinese Law and Government, 2009
O'Dowd, Chinese Military Strategy, Routledge, 2007
Zhang, Deng Xiaoping's Long War, UNC Press, 2015
Zhang, China's 1979 War, China Quarterly 184, 2005
Chen, China's War with Vietnam, Hoover Institution Press, 1987
WarHistory.org, Reassessing the Sino-Vietnamese conflict, 2019
Fravel, review of Zhang, H-Diplo Roundtable XVII-22, 2016
Xia, introduction, H-Diplo Roundtable XVII-22, 2016
Li, A History of the Modern Chinese Army, 2007
Grossheim, Journal of Southeast Asian Studies 53:3, 2022
Yin and Path, Remembering and Forgetting, TRaNS 9:1, 2021
Diamant and O'Brien, Veterans' Political Activism, Modern China, 2015
Foreign Relations of the United States, 1977-1980, volume XIII
Wain, The Indochina Refugee Crisis, Foreign Affairs, 1979
Vietnam News Agency, Northern border defence war, 2019
Nguyen Duc Huy, Vi Xuyen war memoir, via VOV, 2022
Tuoi Tre, Vi Xuyen remains and veterans, 2020
VietnamNet, Division 356 at Vi Xuyen, 2017
Bao Chinh Phu, general secretary at Vi Xuyen, 2025
VietnamPlus, martyrs' remains campaign, 2026
Radio Free Asia, veterans at Pingxiang, 2014 and 2017
The Diplomat, How the war was forgotten, 2023
The Leaflet publishes a Cold File twice a week. Follow at t.me/TheLeaflet for the rest of the series. More Cold Files: Congo War Death Toll and Operation Menu Cambodia.
Next in Cold Files: Algeria's disappeared, 1992 to 1998. A commission that was forbidden to name anyone, a number that runs from 6,146 to 20,000 depending on who is counting, and families still being arrested at weekly vigils in 2026.
Comments