The numbers game: counting civilian deaths in Sri Lanka’s war

Estimating the number of civilian deaths in the final stages of the war in Sri Lanka is proving problematic.

By Kath Noble

The generation-long war in Sri Lanka came to an end in May 2009, with the military defeat of the the Liberation Tigers of Tamil Eelam (LTTE) by government forces. Tamil diaspora groups claimed there had been genocide, but the dominant narrative was of a bloody but essentially fair fight, as captured in the congratulatory resolution passed in the UN Human Rights Council barely a week later.

Even the United States, which backed an alternative and more critical statement, privately felt the same way—a cable published by Wikileaks quotes its Ambassador at Large for War Crimes Issues as having said at around the same time, ‘The Army could have won the military battle faster with higher civilian casualties, yet chose a slower approach which led to a greater number of Sri Lankan military deaths.’

However, this near-consensus has gradually been eroded, and pressure is now mounting for an international investigation.

Although there have been a number of revelations of extrajudicial killings—including the apparent murder of the young son of LTTE chief Prabhakaran—the momentum behind what is an increasingly popular global campaign stems from estimates of the number of civilian deaths in the last stages of the fighting, which have gone up considerably.

The figure most commonly cited in the press was once 7000. Now it is 40 000, 70 000 or even 147 000.

However, this inflation is not the result of any new information.

A recent study entitled The numbers game by a Sri Lankan expatriate writing for personal reasons under the pseudonym Citizen Silva has made an important contribution to the debate on what happened in his country by explaining where these estimates come from, questioning the assumption generally made by journalists that we cannot come to any conclusions without an international investigation.

The 7000 is based on specific casualty reports, as recorded by a network of informants set up by the UN in January 2009. This included more than 200 of its local staff and the local staff of international NGOs—who had been prevented by the LTTE from leaving with their foreign colleagues in October 2008—plus various medical officers, government agents, clergy, education department staff and community leaders.

They compiled reports from around the Vanni for the purpose of keeping the international community informed of the ground situation. Their figures were leaked at the time—17 810 civilian deaths up to 13 May, of which 7737 had been verified by more than one source. (Verification was considered important to account for the pressure that was being brought to bear on the informants by the LTTE, which was keen to present as appalling a picture as possible so as to provoke an intervention under the doctrine of ‘Responsibility to protect’.)

After 13 May, this monitoring became impossible due to the intensity of the fighting. The UN extrapolated on the basis of what it believed to be the daily body count to reach a total of around 11 400 civilian deaths in five months, and later on the basis of information about a single incident increased its estimate to 20 000.

At the time, the United States thought this was an exaggeration. The State Department said in its Report to Congress on incidents during the recent conflict in Sri Lanka, ‘The UN did not rigorously seek to exclude the deaths of possible LTTE conscripts.’

Nevertheless, the UN has since decided that estimates of 40 000 or 70 000 are credible.

These figures—and the still higher one of 147 000—are based on the number of people supposedly unaccounted for in the Vanni.

The 40 000 corresponds to 330 000 minus 290 000, or the population in the No-Fire Zone at the end of February 2009, according to an Assistant Government Agent by the name of Parthipan, minus the number of internally displaced persons (IDPs) who had been registered by the government in collaboration with the UN Office for the Coordination of Humanitarian Affairs by the end of May 2009.

The 70 000 corresponds to 360 000 minus 290 000, or the number of people in the Vanni in January 2009, according to Government Agent for Mullaitivu, Imelda Sukumar, minus the number of IDPs at the end of the conflict.

And the 147 000 corresponds to 429 000 minus 282 000, or the population of the Vanni in October 2008, according to the district offices of Mullaitivu and Kilinochchi, minus the number of IDPs in July 2009.

The report points out that the 290 000 IDPs were not the only people to come out of the Vanni. In May 2009, another 12 000 people were being held by the government on suspicion of being LTTE cadres, while an unknown number paid to escape the camps.

More crucially, it exposes serious discrepancies in the population figures by comparing them.

It notes that there is a very obvious problem with the figure for October 2008. If it is accurate, 69 000 people had vanished into thin air by January 2009.

Also, the same Assistant Government Agent Parthipan who estimated the population in the No-Fire Zone at the end of February 2009 as 330 000 said that it was 305 000 at the end of March 2009 and 150 000 at the end of April 2009.

Meanwhile, the number of registered IDPs had increased from 36 000 to 57 000 and 172 000, implying that 4000 people went missing in March and 40 000 in April. If such huge numbers had been killed, this would have been captured by the informants’ network. Even TamilNet—a website associated with the LTTE—claimed only 2600 civilian deaths in April and 1700 in March.

The report notes that the population figures were put together from lists maintained by village leaders, whose involvement in inflating the numbers for their own private gain or to serve the LTTE agenda has long been accepted. It also points out that while the first organisation to suggest a total of 40 000 stated that even if the counts had been conducted in good faith, they definitely included LTTE cadres, none of the other individuals or agencies who have used the same method have taken this into account.

A local human rights group has claimed that the LTTE maintained a strength of 15 000 until the very end by means of absolutely ruthless conscription. They have said, ‘They were conscripted and used briefly like disposable objects, were brought by the dozens, about 50 a day on average, on trailers of tractors and buried unceremoniously, about three in the same hole, one above the other, covered and forgotten.’

In short, The numbers game categorically denies the credibility of any estimate arising from the second method.

It acknowledges that the first method, too, is flawed, since at least some people may have died without being seen or without their death being recorded by the network of informants—not all bodies would have been transported to medical facilities.

To get over this problem, The numbers game works from data on the number of injuries, on the basis that the injured would have all sought help.

The report starts by estimating ratios of the number of deaths to the number of injuries during the various stages of the final phase. The calculations begin from 20 January, when the government first declared a no-fire zone, and it is assumed that the situation became worse as time went on until the end of the conflict on 19 May, in particular with the Army’s incursion into the No-Fire Zone on 20 April  and from 9 May as the operation to capture the No-Fire Zone commenced.

It uses a ratio of dead to injured from serious injuries of between 40 per cent and 50 per cent —worse than the 33 per cent to 50 per cent range that was accepted by the Panel of Experts appointed by the UN Secretary General—and a ratio of between 20 per cent and 40 per cent for lesser injuries, based on the findings of a local human rights group from their interviews with eyewitnesses. The resulting average ratios of dead to injured range from 60 per cent in January to 90 per cent in May.

These estimates are then combined with another conclusion of the local human rights group, that at least 50 per cent of the injured were shipped out of the No-Fire Zone by the International Committee of the Red Cross (ICRC) between 10 February and 9 May, to give the totals shown in the table. (Figures for the periods up to 10 February and after 9 May were calculated by comparing eyewitness accounts from various sources.)

The total body count is around 15 000.

The approach is checked by means of three case studies—a report by the head of an NGO connected to the LTTE about shelling on 9 and 10 March, the Army’s incursion into the No-Fire Zone on 20 April, as described by the TamilNet correspondent and staff of the ICRC, and interviews by the international media of local doctors when a mortar hit the admissions ward of a makeshift hospital on 12 May. In all of them, calculating the number of deaths from the number of injuries results in exaggeration.





20–31 January



1–28 February



1–31 March



1–19 April



20–30 April



1–8 May



9–14 May



15–16 May



17–18 May




21 500

15 500

When compared to other estimates of the total body count at various points in the final phase, this method also comes up with higher figures.

Unfortunately, The numbers game has been ignored by almost all media.

It is easy and comfortable for journalists to call for an international investigation to settle the matter, but making this happen is not so simple. After three decades of fighting, opinion in Sri Lanka is deeply polarised. No intervention will be automatically regarded as impartial by the majority of all communities—any comment on what happened in May 2009 has to prove its worth by engaging with serious analysis that presents a contrary view. Failing to do so also risks contributing to the further polarisation of society.

Kath Noble is a journalist. She worked in Sri Lanka from 2004 to 2010, and has since been engaged in postgraduate studies at Jawaharlal Nehru University in India, focusing on the political economy of the post-war reconstruction in Sri Lanka. She may be contacted at


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s