Sri Lanka’s Dead and Missing

The need for an accounting

by International Crisis Group blog, February 27, 2012

A demonstrator cries while holding a picture of her relative who went missing during Sri Lanka's war with the Liberation Tigers of Tamil Eelam (LTTE) as she takes part in a protest in Colombo September 9, 2009. Photo: Andrew Caballero-Reynolds/Reuters

Nearly three years since the end of the war, there’s a growing need for an accounting of – and for – those killed and missing in the final months of fighting in northern Sri Lanka in 2009. Members of the UN Human Rights Council, opening its 19th session in Geneva today, should be ready to press the Sri Lankan government for real answers.

Instead of grappling with the many credible sources of information suggesting tens of thousands of civilians were killed between January and May 2009 – including the UN’s real-time data collection, international satellite imagery, and the government’s own population figures – the government is rewriting history on its own terms. In the lead up to the Human Rights Council session, the government released an “Enumeration of Vital Events” for the Northern Province. It finds the total death toll during the five bloody months of fighting in 2009 to be under 7,000 with another 2,500 missing, but it doesn’t differentiate between civilians and combatants or assign responsibility for any death to either the Liberation Tigers of Tamil Eelam (LTTE) or to government forces.

These findings fall far short of the UN Secretary-General’s panel of experts’ estimate that as many as 40,000 civilians died in those last months and even higher estimates based on the government’s own prior census figures. This “enumeration” also runs counter to an important recommendation of the government’s Lessons Learnt and Reconciliation Commission (LLRC) calling on the government to “conduct a professionally designed household survey covering all affected families in all parts of the island to ascertain first-hand the scale and the circumstances of death and injury to civilians, as well as damage to property during the period of the conflict”. The LLRC made this recommendation at the end of November 2011, well after the “enumeration” project was conducted from June to August 2011.

The LLRC’s report has serious shortcomings in its treatment of allegations of war crimes and crimes against humanity by government forces, but it also acknowledges important realities, including breaking with years of government claims of “zero civilian casualties” and accepting that “considerable civilian casualties had in fact occurred during the final phase of the conflict”. While it then goes to lengths to absolve the government of responsibility for those casualties, its recommendation of a professionally designed survey could help clarify the fate of the dead and missing in the north – if done credibly and independently. Unfortunately, there are no signs of such a credible and independent process emerging.

Just days after the LLRC delivered its report to the president on 20 November, his brother, Defence Secretary Gotabaya Rajapaksa, delivered a speech to the “Inaugural National Conference on Reconciliation” in which he said that the government had “conducted a complete census” in the north, which was “in the process of finalisation”. He said a “questionnaire” had been used which included seven categories of dead or missing persons: “those who died of natural causes; those who died of accidents; those who left th[e] country through illegal means, particularly by boat to India or to South East Asia, and from there to the West; those who died whilst fighting as members of the LTTE; those who died as a result of being coerced to fight by the LTTE; those who died as a result of resisting the LTTE … ; [and deaths] that occurred due to military action”.

Without explaining how those categories were determined or what the government had done to avoid bias in its questionnaire and information gathering, the defence secretary concluded that “as a result of the census, we already know that the real number of dead and missing is far too small to provide any substance to the absurd allegations of genocide and war crimes that have been made against our military by the rump LTTE and their cronies”. On 8 February at an event hosted by the Swiss mission to the UN in New York, Sri Lanka’s permanent representative to the UN repeated the claim that the government has carried out “a comprehensive census in the Northern Province which will enable firm and verifiable conclusions to be derived at on issues involving disappearances, deaths, etc”.

The “enumeration” released soon after this statement doesn’t deliver the details promised. Instead it appears to be little more than another attempt to short-circuit what Sri Lanka’s people really need and what the global community – and even the LLRC – have been asking for: an independent, credible assessment of who and how many died and under what circumstances (i.e., who and what caused the deaths or disappearances).

The process by which this supposed count has happened is not at all clear, but the conditions under which any census would have been conducted in the north – an area under effective military occupation by a victorious army accused of the crimes in question – are not conducive to a fair and accurate count. The risk that the government has now politicised the department of census and statistics, as it already has politicised the police, judiciary and human rights commission, should be deeply worrying to Sri Lanka’s international partners.

As member states of the Human Rights Council prepare for the upcoming session, they should ask the government for a full explanation of how this purported census was conducted, what safeguards were in place to ensure independence, who (by name and by civilian or combatant) was killed or went missing and how, and whether UN agencies and independent civil society organisations will be allowed to verify the findings. Equally important, Council members should ask the government to reconcile its “enumeration” with the now-extensive information available suggesting that tens of thousands of civilians were killed in the final stages of the war, most as a result of government fire into heavily populated civilian areas.

The precise number and identities of all civilians killed in those last bloody months will likely never be known, especially if the government persists in its resistance to an independent, international inquiry. Nonetheless, several sources of information regarding civilian casualties need to be addressed – fully and transparently.

Sources of information suggesting civilian deaths in the tens of thousands

The UN’s real-time, on the ground survey of deaths and injuries

There are, first of all, the casualty figures gathered by the UN and humanitarian staff trapped in the fighting, which recorded 7,721 civilians killed and 18,479 injured between August 2008 and 13 May 2009, after which information collection became too difficult. These numbers were not estimates, but actual counts based on eyewitness sightings verified by two additional sources. The vast majority of those included in the UN count were killed between late January and late April 2009, before the escalation in fighting in the final three weeks.

The Secretary-General’s panel of experts noted strong grounds to believe these numbers understate actual casualties during that period. These include the conservative methodology used to collect the figures, suspected underreporting by UN agencies (in response to pressure from the Sri Lankan government), the location of many casualties in areas inaccessible to observers, and the fact that following 13 May, the number of civilian casualties likely increased significantly as many civilians died from their injuries with no functioning hospital or humanitarian facilities in the warzone to register casualties or treat the wounded.

The government, including the LLRC, has frequently dismissed these figures because the UN failed to publish them, without acknowledging that the UN raised them with the government in private discussions. The UN’s failure to speak out about its casualty estimates and the scale of the humanitarian crisis, and the government’s willingness to castigate the UN for even attempting to assess the civilian toll, are both matters of serious continuing concern.

Casualty estimates for the final week of fighting

Multiple eyewitnesses who were in the warzone during the final week estimate that thousands of civilians were killed in those days. Many describe walking past hundreds of bodies as they exited the final “no-fire zone” and seeing entire families buried in collapsed bunkers. A U.S. embassy cable on 18 May 2009, the day the government declared the war over, said a UN contact thought the LTTE’s claims of 25,000 civilians wounded or killed in the last few days were exaggerated, but that based on a 10 May shelter analysis and estimates of 70,000 to 80,000 people in the “no-fire zone” before the final assault, the number unaccounted for could be as high as 7,000 to 17,000. The UN contact also reportedly said the UN doubted the LTTE’s claims on the night of 17 May that it still had 1,000 to 2,000 cadres.

Contemporary population figures from senior government officials

There is also information from the government’s own officials working in the north suggesting that as many as 70,000 or even 140,000 civilians who were surveyed in the warzone just before or during the final months of fighting never made it to the government internment camps at the end of the war.

At least three separate figures need to be compared against the number of civilians in the camps as of late May 2009, which the government reported to be approximately 290,000. First, the former district secretary for Mullaitivu and current district secretary for Jaffna, Imelda Sukumar, testified to the LLRC on 4 November 2010 that there were 360,000 people caught in the fighting in the Puttumatalan “no-fire zone” established in February 2009.

Second, the UN panel of experts reported that her assistant, the former additional government agent (AGA) of Mullaitivu, and his staff who were in the “no-fire zone” counted some 330,000 people still trapped in the fighting in early February 2009. At that time, government figures showed that 35,000 were already in government camps. After the AGA advised officials in Colombo of the 330,000 figure, they wrote to him that the figure was “arbitrary and baseless” and that the government would be “reluctantly compelled” to discipline him for providing “wrong information to any source especially in regard to IDP figures”.

Finally, documents from the local government offices in Kilinochchi and Mullaitivu districts, dated 30 September and 1 October 2008, available here and here, show a total population of 429,000. These figures were cited in the LLRC testimony of the Catholic Bishop of Mannar, who asked for clarification as to what happened to the more than 140,000 people apparently missing given the much smaller population corralled into government camps.

Corroborating the government’s contemporary population figures

This last estimate of the number of those still unaccounted for may seem high. Indeed, the central government has long argued that local population figures were inflated under pressure from the LTTE, in order to exaggerate the humanitarian crisis and to generate greater quantities of humanitarian supplies, which the LTTE could steal. While some inflation in the figures is possible, it is unlikely to have been large enough to explain all, or even most, of the discrepancies. There is also some corroborating evidence that argues for taking seriously even large estimates of the missing and demanding a full and independent accounting.

For example, if one takes the total population figures for residents of Kilinochchi and Mullaitivu districts from the 30 September and 1 October 2008 local government documents, they match almost exactly the central government’s estimated 2008 population figures for those districts, which can still be found on the department of census and statistics website in its estimated mid-year population by sex and district, 2000-2010.

The combined total of Kilinochchi residents counted by local officials in late 2008 was 146,700 (121,900 then living in Kilinochchi and 24,800 displaced to Mullaitivu), while the central government estimate for Kilinochchi residents was 140,000 in 2006, 143,000 in 2007 and 147,000 in 2008 (and provisionally 154,000 in 2009 and 156,000 in 2010).

Similarly, the combined total for Mullaitivu residents counted by local officials in late 2008 came to 126,350 (100,600 still living in Mullaitivu and 25,745 displaced to Kilinochchi), while the central government figures for Mullaitivu show 129,000 in 2006, 132,000 in 2007 and 135,000 in 2008 (and provisionally 154,000 in 2009 and 148,000 in 2010).

Because local government officials’ figures for Kilinochchi and Mullaitivu residents in late 2008 are almost exactly what the government had officially accepted for years, the government should explain why so many fewer people ended up in government camps in mid-2009. It should also explain why its most recent “enumeration”, which says there are now only 103,717 people in Kilinochchi and only 66,526 people in Mullaitivu – a drop of over 100,000, doesn’t raise many more questions than it answers.

A UN survey in the internment camps

Another intriguing statistic can be found in a UN Survey of 100 randomly selected shelters in zone 3 of Menik Farm in early May 2009 – prior to the worst fighting of the final two weeks. This small survey found that “22 per cent of the families” interviewed “reported that an immediate family member had died”. Extending this percentage to the approximately 90,000 families who ended up in camps after the end of the war, it suggests a minimum of 18,000 killed. Sample bias and other potential problems with this survey – including the possible inclusion of some combatants or deaths in earlier stages of the war – need to be examined, yet given the timeframe (prior to the deadliest weeks) and the possibility of multiple deaths within (or complete destruction of) some families, it could in fact be a generous minimum.

Estimates of war widows and female-headed households

Finally, there are other estimates available – including from the government – that appear to be consistent with large-scale loss of life. For example, multiple sources have claimed that there are now 40,000 “war widows” in the north. In September 2010, the ministry for child development and women’s affairs said it had lists of 40,000 war widows in the north, though it reduced this number without explanation in August 2011, to only 16,936. A separate mediareport cites government and donor figures of 30,000 out of 110,000 households in the former warzone that are headed by women. And a survey by the Jaffna-based Center for Women and Development reportedly estimated 40,000 female-headed households in the north, half of those in Jaffna. Not all of these women would have lost their husbands in the final months of the war, nor would all those men killed have been civilians – but many would have. And to the number of non-combatant husbands killed, one would have to add the women, children and unmarried men who died in the fighting, as well as those cases where both members of a married couple – and even whole families – were killed.

What the LLRC said about civilian casualties

The LLRC reported that the scale of civilian casualties, especially from January to May 2009, was a key question for the commission. Yet it accepted what the defence ministry told it – that “an estimate of civilian deaths was not available”. At the same time, the ministry had no problem providing an estimate of LTTE deaths – 22,247 for July 2006 to May 2009, with 4,264 confirmed by name for the period January to May 2009; or an estimate of security force deaths – 5,556 for July 2006 to May 2009.

Separately, the LLRC noted that the defence ministry had estimated the total number of LTTE cadres in the north to be 21,500. Given that approximately 11,700 suspected cadres were detained for “rehabilitation” at the end of the fighting, there are serious questions as to (1) how the government reconciles its 21,500 cadre estimate with its total of 34,000 killed or detained; (2) whether the 22,247 LTTE deaths were combatant deaths; and (2) whether the 11,700 detained for “rehabilitation” were in fact combatants. Unfortunately, the LLRC did not acknowledge, let alone answer, any of these questions.

Instead, the defence ministry told the LLRC that “it would be extremely difficult, if not impossible, to distinguish between LTTE and civilian casualties”. The LLRC expressed its “regret” at the absence of any official record or post-conflict estimate of civilian casualties while at the same time concluding that “considerable civilian casualties had in fact occurred”, but placed the blame primarily on unexplained “crossfire” and on the LTTE – just as the government and military officials who testified before it did.

The LLRC’s recommendation of a professionally designed island-wide household survey regarding civilian deaths and injuries – if done independently and credibly – could make up in part for its unwillingness to challenge the government’s narrative. Such a survey could also provide all Sri Lankans more clarity regarding how many and whose lives were lost in the war, including thousands of missing soldiers and people killed or disappeared outside of the north through decades of counter insurgency operations.