NATO and the Pentagon have refused to estimate the civilian death toll in last year's 78-day bombing of Yugoslavia, calling such a calculation beyond their powers. On Sunday, Feb. 6, a study by New York-based Human Rights Watch (HRW) confidently placed the number of civilian deaths at around 500. How did HRW arrive at that number?
During the NATO bombing, HRW monitors recorded the location and destructiveness of each attack based on information gleaned from Yugoslav media, online dispatches by Yugoslav official agencies, and e-mail correspondence with Yugoslav officials. HRW crosschecked this list against accounts from the media, government, and military in NATO countries.
After the bombing, HRW dispatched body-count investigators on a three-week mission to Yugoslavia. Once there, they:
1) Physically inspected sites. The HRW teams visited 42 of the 90 sites confirmed as "incident locations." The highest priority was given to targets thought questionable under international humanitarian law, such as demolished power plants, bridges, and radio and television stations. Lower priority was given to sites where only one or two civilians died and for which consistent press and official reports were available. Some sites were excluded because they were either inaccessible or completely leveled.
HRW teams inspected sites to judge whether NATO bombs or Serb initiative caused casualties. For example, smoke stains up and down a window indicated a building had been burned from within, and put the likely onus on the Serbs. The Serbs were also fingered when scattered "points of destruction" were found, as NATO bombs would have landed in a distinctive, straight-line formation. In one case, in which Serb officials blamed the death of 95 imprisoned civilians on NATO, HRW attributed 25 deaths to NATO and the other 70 to Serb paramilitaries.
2) Inspected official government documents. These documents included autopsy reports, death certificates, civil-defense reports, and photographic evidence of bomb damage. While Serb propaganda stubbornly placed the number of civilian deaths between 1,200 and 5,000, information extracted from some 900 pages of autopsy reports helps confirm the 500-person figure favored by HRW.
3) Conducted interviews. HRW interviewed eyewitnesses, neighbors and relatives of the deceased, local reporters, firefighters, rescue workers, and factory and utility-company representatives. HRW typically used these interviews to double-check government documents.
Of the 90 incidents confirmed by HRW (42 by physical investigation, the rest with documents and interviews), the organization claims to have identified the precise number of victims (by name) in 69 incidents. In seven incidents it claims to have determined the precise number killed, but only some of the names. In 11 incidents it determined precise numbers but no names. Neither names nor numbers killed could be established for the final three incidents, so HRW estimated a range of 0 to 13 deaths in each incident, producing a range of 488 to 527 total civilian deaths.
Explainer thanks Bill Arkin, military consultant for Human Rights Watch.