Encountering 'Death Counts' in the Final Phase of the Eelam War


by G. H. Peiris

1. Preamble

Almost all recent media reports in the west on the impending United States resolution against Sri Lanka at the Human Rights Council 'Periodic Review' scheduled for next March state as a firmly established fact that 40,000 civilians were killed in the course of the final phase of the Eelam War. Most of these reports also imply that, in sponsoring such a resolution, US government is impelled by its conviction that the government of Sri Lanka is accountable for that colossal crime. In none of these reports is there a reference to the fact that even the Darusman Report, despite the blatant duplicity of its approach, and replete though it is with distortions, has never said that 40,000 Vanni civilians were killed during that phase of the war. What it did say was that the number of civilian deaths could be as high as 40,000.

But to the United States and its allies in Europe that are arrayed against post-war Sri Lanka trivia such as factual accuracy and fairness do not seem to matter. There is sufficient evidence in the confidential communications from their own official informants - diplomatic missions stationed in Colombo - that the Sri Lanka security forces exercised greater caution to avoid civilian deaths than their own armies have ever done in military offensives conducted outside the United States, and even there if one were to think of the 19th century. That appears to be of no consequence. What matters is 'regime change' - the need to evict the popularly elected government of Sri Lanka because it does not genuflect in the way minions of the international community are required to do, and a charge of 'war crimes' and 'crimes against humanity' is the only façade that could conceal the clandestine strategies being adopted to achieve that cardinal objective.

How the self-appointed global guardians of human rights help sustain the crumbling hegemonic global power arrangement is vividly illustrated in the media account of the ceremony held in London for the launch of Frances Harrison's Still Counting the Dead (October 2012). Here are some extracts from the pronouncement reported to have been made on that occasion.

Eric Solheim - Norwegian politician who, in his role as 'mediator' in the Sri Lankan conflict, hardly ever bothered to conceal his commitment to promote LTTE interests during the last 10 years of the Eelam War.

"In Sri Lanka, the government was winning the war and victory was at hand. They had no intention of stopping. In January 2009, the government declared war. The call for the LTTE to accept an organized end to the war which included the LTTE handing over weapons, registering LTTE cardres and every single Tamil civilian supervised by international authorities - the UN, US, India, etc. was not heeded … If that happened, the lives of all the Tamil Tigers and Tamil civilians could have been saved. But, Prabhakaran rejected this offer. If the LTTE agreed, the Sri Lankan government had no other option than to accept it. Though the LTTE leadership rejected the offer, it can never ever be used as an excuse by the government to indiscriminately bomb very dense Tamil areas".

Alan Keenan - International Crisis Group's Sri Lanka Project Director, according to whom 40,000 to 147,000 civilians were killed in the Vanni war.

"If you look at the other conflicts around the world where there have been many thousands of people killed in illegal ways, they don't generally get sorted out legally very quickly. There is a long game to be played, a frustrating game for all involved specially families of the victims, survivors, but I think, over the longer term, there are still options, there are still possibilities. What we need are those involved to produce documentaries like Channel 4, books like Frances Harrison's, Reports like my organization (International Crisis Group) can continue to produce....Together, cumulatively, with effective lobbying, we can keep the Rajapaksa regime and those associated with them on the defensive and ultimately, I think, as political developments happen within Sri Lanka and outside, there will be a moment when they are not reigning supreme as they are now, then these issues come back to bite them".

Yasmin Sooka, One of the 3-member UNSG 'Panel of Experts on Accountability in Sri Lanka' and co-author of the Panel report which placed the upper margin of civilian deaths in the Vanni war zone at 40,000.

The government of Sri Lanka claims that they mounted a humanitarian mission to rescue the civilians from the LTTE. GOSL also claim that there were zero civilian casualties. "But we think that as many as 70,000 civilians died. I think LLRC did produce some valuable lessons but in terms of the question of accountability it is a big failure".

Callum Macrae, producer of the macabre Channel 4 documentary, 'Sri Lanka's Killing Fields'.

(His fanaticism seems to make him forget that crude melodrama like his are produced by the hundreds at Hollywood, Bollywood, Kollywood and many other places all over the world).

"We need to keep telling this story. Most importantly, we want to get this film by the end of January and launch it internationally in the build-up to the Human Rights Council. We want to show this film in Asian countries and African countries as well. We need to go to the Non Aligned nations and say this is a matter of human rights. This is a fundamental matter of international law. This is an issue the rest of the world has to take up. This will be ready for the Human Rights Council. Then we need to use this film to campaign to CHOGM at the end of the year in Sri Lanka".

Frances Harrison, BBC Correspondent who has had close contact with the Tiger leadership from about 2002 is the author of Still Counting the Dead. Reproduced below is only a collection of a few brief extracts - not the really spicy flights of imagination.

"Today marks three years since the end of the fighting in Sri Lanka. I would like to mourn the dead but still I do not know how many. Estimates range from 7 to 147,000. It is a shocking difference. How is it possible in this world of satellites, rolling news and internet we have no idea how many human beings really perished, even rounded up to the nearest thousand? … Numbed to the sight of death, families were forced to abandon the corpses of their loved ones as they ran for their lives. A brave doctor who saved thousands of lives is haunted by the memory of the 150 patients he abandoned under a tree on the last day of the war; he can no longer stand the sight of blood and does not want to be a surgeon. It was a place where loving parents discussed suicide with their children, unable to tolerate the agony of dying one by one. A medic saw a baby born with a bullet lodged in his tiny leg, shot while still in the womb. In the makeshift hospitals dying mothers screamed for their babies to give them one last feed - knowing the breast milk would be their last gift of life".

To those of the west, mesmerized by this type of 'truth', what do numbers matter? Why count the dead even if you could devise a reasonably acceptable method of doing so, unless you just go on counting until you somehow reach a total which the Sookas, the Harrisons and the Macraes of this world say now it's sufficient for our purpose? There are tens of thousands in Sri Lanka who mourn the dead - relatives and friends - but not with crocodile tears and not in pursuit of money, power and fame. We are intensely saddened every time we learn about the murder of countless civilians in, say, Vietnam, Cambodia, Libya, Iraq, Pakistan or Afghanistan - a process that will go on and on until the "world is safe for democracy". We are also aware that, for certain mourners, the Vanni dead are very special (if not unique) in that there is plenty of diaspora largesse for those who invent death-counts or bemoan the absence of a count.

2. A Cross-Section of Estimates

I have examined seven important studies that focus (some of them, not exclusively) on the subject of civilian deaths in the parts of the Vanni lowlands that were under the control of the LTTE from early January to 19 May, 2009. They are (arranged in chronological order) the following:

UNIVERSITY TEACHERS FOR HUMAN RIGHTS-JAFFNA/UTHR-J (2009) Special Report No. 32 of 10 June 2009 and Special Report No. 34 of 13 December 2009

DARUSMAN, Marzuki, SOOKA, Yasmin & RATNER, Steven R. (2011) Report of the Secretary General's Panel of Experts on Accountability in Sri Lanka, a.k.a. Darusman Report'/UNSG-PoE Report

GOVERNMENT OF SRI LANKA (November 2011) Report of the Commission of Inquiry on Lessons Learnt and Reconciliation, www.slembassyusa.org/downloads/LLRC-REPORT.pdf

DEPARTMENT OF CENSUS & STATISTICS/DCS, SRI LANKA (2012) Enumeration of Vital Events, Ministry of Finance & Planning, Colombo


ROBERTS, Michael (2013) Estimates of the Tamil Civilian Death Toll during the Last Phase of Eelam War IV,

ENGAGE SRI LANKA (2013) Corrupted Journalism, Channel 4 and Sri Lanka, www.engagesrilanka.com

In addition, there are many shorter reports such as those by the United Nations Country Team (2009); Human Rights Watch/HRW (2009) Displaced and Detained: The Plight of Civilians in Sri Lanka's Vanni Region; Report to the Congress on Incidents During the Recent Conflict in Sri Lanka, submitted by the United States, Department of State (2009); records of 'monitoring' the conflict by the United Nations Office for Coordination of Humanitarian Affairs/UN-OCHA, Colombo; and similar documents published by the International Committee of the Red Cross/ICRC, Geneva. To these may be added the articles by reputed journalists like Reddy (2009) 'Final Hours - and Eyewitness Account of the last 70 hours of Eelam War IV,' that deserve serious attention.

It is almost universally admitted that there could be no exactitude in estimating the number of "non-combatants"/"civilians"/"Internally Displaced Persons - IDPs" entrapped by the LTTE in the course of its retreat across the eastern Vanni, and what proportion of that number lost their lives having been caught in the cross-fire ("collateral mortalities") or, as frequently alleged, wilfully targeted by one or the other of the warring parties. The related gross estimates (some of which are tabulated below) vary from less than 2,000 to well over 145,000. There is also considerable diversity in the methods followed in the different computations. While some estimates are no more than components of the rabid post-war hostilities directed at Sri Lanka under the pretext of articulating humanitarian concerns, others are based at least on a semblance of rational 'methodology' in the analysis of parameters relevant to the arrival at rough but plausible approximations. Even among these there are the products of agents provocateurs impelled by either sanctimonious prejudices or hegemonic geopolitical objectives.


3. Computation Methods

The methods employed by the authors/institutions referred to above in their attempts to estimate the Vanni death toll during the final phase of the Eelam War could be placed in the following categories:

Population Change Method - This has entailed the assumption that the difference between the estimated Vanni population at the commencement of the final phase and the total number of persons at the end of that phase (considered as being more accurately recorded at IDP camps, interment centres and hospitals etc.) provides an acceptable estimate of the number that died and the number 'missing'.

Injury-to-Death Ratio Method - This method has entailed the use of (a) the supposedly comprehensive data maintained at hospitals in the Eastern Vanni on treatment of the wounded, and (b) the scatter of information available on the number that died and the number wounded (over spells of fighting), for the purpose of working out a set of 'injury-to-death' ratios for different time-spans of the Final Phase, and then, the projection of such ratios over the entire period to generate an overall death-count.

Sporadic Information Method - This methods has entailed the use of a large bodies of "credible evidence" obtained from miscellaneous sources such as eyewitness accounts, writings by journalists, reports compiled by monitoring institutions outside the war zone (those such as the UTHR-J, TamilNet and the ICRC) which claim to have had networks of informants within the war zone, and, in a few instances, demographic data claimed to have been extracted from satellite images in the course of the war, to reach what those who have employed this method refer to as "credible estimates".

Satellite Imagery Interpretation Method - Satellite images are said to facilitate actual headcounts of the war zone population at various stages of the Final Phase, and (somewhat dubiously) to enable the making of a distinction between combatants and non-combatants in the process of counting. The impact of bombing and shelling on war zone targets and the placements and directions of fire of the heavy artillery used by the Sri Lanka army are also claimed to be revealed by satellite imagery.

Post-Conflict Census Enumerations - Household headcounts conducted by government agencies designed mainly to elicit information on the war casualties obtained surviving from kith and kin.

In their application to the Vanni war these have not been mutually exclusive methods. Most of the estimates referred to in Table 1 have involved the use of more than one of these methods, the 'methodological' differences among them being mainly in the relative emphasis placed on one or another of the methods.

3.1. 'Population Change Method': Applications

A ballpark estimate relating to the Final Phase of the war, which has remained undisputed at least in serious writings is that the number of people who survived the war, having escaped from the clutches of the LTTE at various times since early January 2009, or been evacuated by the ICRC-sponsored 'mercy missions', or captured by or surrendered to the Sri Lanka Army ranged between 295,000 and 310,000. Accordingly, if it is possible to identify a plausible estimate of the population in the LTTE-controlled area at the commencement of the Final Phase (from among those available - see the list presented below), the difference between the latter and the former number could be considered an acceptable estimate of roughly how many of the Vanni population perished (or considered "missing") during this 139-day period.

Imelda Sukumar (Assistant Government Agent, Mullaitivu) estimated that the Vanni population by about January 20, 2009 was about 360,000 [Source: Independent Diaspora Analysis Group (IDAG Report, p. 20)

K. Parthipan (Assistant Government Agent, working in the LTTE-controlled area) reported on 3 March 2009 that the population of the area on 3 March 2009 was 330,000 and, on 29 April, 150,000.

Human Rights Watch (an international non-government organisation), placed the estimate on civilians in the Vanni in December 2008 at "between 230,000 and 300,000". (Source: 'Sri Lanka - Trapped and Mistreated: LTTE Abuses against Civilians in the Vanni' 15 December 2008)

Time news magazine (of February 3, 2009) reported that "as many as 250,000 civilians" were being held hostage by the LTTE .

UN Country Team (report on the situation at the end of February 2009) 269,000 in the LTTE-controlled area on 28 February, and 36,000 on 25 February in IDP camps

UN 'World Food Programme' (WFP) estimated in February 2009 that the population in the LTTE controlled area at that time was 250,000 ( http://www.wfp.org/stories/sri-lanka-vanni)

University Teachers for Human Rights - Jaffna (UTHR -J), Special Report No. 34 of 13December 2006, Sections 5.1. & 5.2. - The UTHR-J placed the "original Vanni population" tally at about 365,000

There are, however, no estimates of the size of the Vanni population in the immediate aftermath of eviction of the LTTE from its headquarters at Kilinochchi on 1-2 January 2009 - The closest available to that point of time is a computation by Ms. Imelda Sukumar, the Government Agent of Kilinochchi District, which, in certain writings, has been considered an acceptable 'base estimate' of the Vanni population in January for the purpose of tracing the subsequent changes.

From 1 January to the time of announcement of the 1st NFZ (20 January), 5 Divisions and 3 Taskforces of the Sri Lanka Army were in cautious pursuit of the retreating Tiger cadres that were herding along tens of thousands of civilians as both a human shield for their combatants and a disincentive to enemy artillery and bomb attacks. The landward warfront at this stage, though gradually shrinking in length, stretched over a distance of more than 70 miles - for the most part, over forested territory. In this chaotic demographic scene, despite the presence of a numerically large army on one side of the warfront (but largely confined to narrow corridors along the A34 and A35 'highways'), and Tiger combatants numbering probably about 10,000 on the other side, it would have been possible for a fair number of civilians, especially those not encumbered by infirmities or household ties, to escape from the war zone and 'vanish' into the wilderness of the adjacent parts of Trincomalee and Vavuniya districts.

It is in the context of these considerations that the Government Agent Imelda Sukumar's population estimate of 360,000 for January 2009 should be looked at, especially in the context of the fact that November 2008 (about 6 weeks before the fall of the Tiger base at Kilinochchi), she had stated in an official communication that there was a population (including all IDPs from adjacent districts as well) of 360,000 "with me" (meaning, "within her area of authority at that time"). If this latter estimate, presumably made at the headquarters of district administration with the assistance of support staff, and on the basis of data extracted in 'Householders' List' maintained by the graama niladhaari (village officers) prior to the war-induced large-scale displacement of the Vanni people from their homes from about May 2008 (when the concerted offensive against the LTTE in the Vanni region began), was a fairly accurate estimation, the actual size of the population in the LTTE-controlled of the eastern Vanni on 21 January could by no stretch of imagination have been anywhere near 360,000. An inference that could be drawn from this deliberate or inadvertent error of the part of Ms. Sukumar is that official computations of the Vanni civilian population made from time to time during February, March and April 2009 by her subordinates (especially the Assistant Government Agent, Mr. K. Parthipan) were featured by similar distortions attributable to similar impulses and compulsions. Thus, what all these considerations point to is that the 'Sukumar-Parthipan' estimates are not plausible enough to factor into equations that could generate an acceptable estimate of Vanni war casualties.

Had there been some source from which a more reliable estimate of the 'Vanni population' in January 2009 could be obtained, calculating the 'Population Change' between January and May 19th would have served as a feasible method of quantifying the Vanni death-toll during the Final Phase of the war. The inescapable fact, however, is that the other computations available, such as those by the 'Human Rights Watch' 'UN Country Team', 'UN World Food Programme' (referred to above) are unacceptable for the reason that, when matched against the numbers that are known to have escaped/rescued from the war zone, they appear as under-enumerations.

In the absence of a plausible estimate of the size of the Vanni civilian population in January 2009, the search for the death-count in the Final Phase of the war by employing the 'Population Change Method' has often tended to involve the arbitrary selection of a preferred 'base-population' value, the preference invariably depending on the size of the death-count which the estimator wants to achieve, from the viewpoint of either the objective with which the estimate is being made and/or personal convictions based on qualitative information including anecdote, gossip, rumour and partisan propaganda. This is illustrated, for instance, by the upward revision of the 'UN Secretary General's Panel of Experts' estimate of "up to 40,000 civilian deaths" (in April 2011) to the 'UN Secretary General's Internal Review Panel' estimate of "up to 70,000 civilian deaths (November 2012) - an obvious retaliation against the refusal by the government of Sri Lanka to agree to a UN-sponsored international investigation into alleged "war crimes".

3.2. 'Injury-to-Death Ratio Method': Applications

The assumption on which the adoption of this method of estimating the Vanni death toll is based is that the number or persons wounded (and treated at hospitals and makeshift medical centres), and the number that died during certain episodes of attack (i.e. bombing raids, artillery shelling, firing with infantry weapons etc.) as reported by persons within the war zone and/or estimated by agencies outside, could provide a set of 'injury-to-death' ratios for different spells of the war from January to May 2009, thus facilitating a plausible estimate of the overall death toll for the entire period. The application of this method for estimating civilian mortalities has also entailed a supplementary assumption that, in the maintenance of records of treatment of the wounded, the medical personnel serving the Vanni population at this time did make a fairly accurate distinction between combatants and non-combatants, thus facilitating a fairly accurate estimation of an overall total of civilian deaths. (How the medical personnel could have done so, except in the case of the aged and small children, baffles imagination.)

The most detailed application of this method available so far is found in Section 4.4. (captioned 'The Probable Truth') of the monograph titled The Numbers Game compiled by a group of researchers calling themselves the 'Independent Diaspora Analysis Group' (IDAG). The "probable truth" arrived at by the IDAG through a seemingly complex method of computation is that the total number of 'Vanni civilian deaths' during the final phase of the war was about 15,000; and that another 2-3,000 of the Vanni population remains unaccounted. Presumably, the latter - 'missing' - have been killed while fleeing from 'NFZ 2', or have surreptitiously crept out of the government-run IDP camps, or (in the case of LTTE combatants who survived) escaped incarceration by the Sri Lanka security forces. Projecting backward from their 'injury-to-death computations' and the reported number sheltered in IDP camps in early June 2009, the IDAG has also estimated that the total population (including LTTE cadres) in the Vanni on 10 January 2009 was between 339,300 and 343,300.

In contrast to the report published by the UNSG-POE (a.k.a. 'Darusman Report'), the IDAG report is, indeed, a valuable (arguably the most valuable) contribution to an understanding of many things that happened during the 'Final Phase' of the Eelam Wars. There is, however, considerable doubt about the validity of the IDAG method of computing the 'injured-to-killed' ratios during different spells of the 'Final Phase' - the pivotal parameter of their overall civilian death-toll estimate.

The essence of the information furnished in Section 4.4. of the IDAG Report could be sketched out as follows:

(a) Based largely on computations by the UN Office for the Coordination of Humanitarian Affairs (UN OCHA), the IDAG has estimated the number wounded and the number that died during different spells from late January to early March 2009, and derived the following approximate ratios of these two sets of numbers (Table 2).


(b) The method followed by the IDAG for estimating the 'injured-to-killed ratios' from early March to 19th May has involved, inter alia, the recorded numbers of deaths and persons injured during 3 spells of armed confrontation during this time, referred to as 'Case-Studies' (Table 3).

Following the presentation of its estimates of deaths and the number wounded in these 3 episodes of attack, the IDAG has referred to several sets of additional information extracted almost entirely from TamilNet reports (supplemented with newspaper articles by the journalist D. B. S. Jeyaraj and 'Special Reports' released by UTHR-J) to reach the conclusion that between 1 May and 18 May, 6,082 civilians were injured, and a total (whether this total includes LTTE cadres in not stated) of 5,418 were killed.


Clarifications of Terminology

'Final Phase' refers to the period between early January and 19 May 2005 (i.e. from the time of eviction of the LTTE from the township of Kilinochchi that had served as their principal base since December 1995until the final battle-field defeat of the LTTE).

The 'Vanni' is a geographical designation for a large part of the northern plains of Sri Lanka. 'Eastern Vanni' refers to the area which was under LTTE control at the commencement of the 'Final Phase'- roughly the territory east of the north-south aligned A9 Highway.

The 'War Zone' during the 'Final Phase' declined in extent from almost the whole of the 'Eastern Vanni' in early January to a tiny coastal strip by mid-May 2009.

The 'No-Fire Zones' (NFZs) announced by the Sri Lanka government were small areas within the 'War Zone' - the 1st NFS declared on 20 January (26 km² located between Theravilkulam and Puthukudiyiruppu; a.k.a. 'The vipuram Safety Zone') and the 2nd NFZ (initially a coastal strip of about 10km² north of Mullaitivu town). The 'no fire' stipulation was ignored by the LTTE in the sense that the zones were used as bases for their armed hostilities.

The phrase 'Vanni Population' as used in this study refers to the inhabitants of the LTTE-controlled area (combatants and non-combatants) during the 'Final Phase' of the war.

To be continued tomorrow

animated gif
Processing Request
Please Wait...