Question on Iraqi civilian casualty trends
The much-blogged-about USAToday story about Iraq civilian casualties dropping to one-third of their July levels may be wrong. The July numbers of 3500 were sourced from the Iraqi Ministry of Health (MIH) and the Baghdad morgue (Medico-Legal Institute, MLI). The widely-reported August number of 973 may be based solely on the MIH.
Update: The Baghdad morgue revised their figures upwards a few days later.
Let’s look at reports for the months of June and July, to clarify the sources.
In June, this UN report noted:
9. The reported number of civilian casualties continued an upward trend. According to figures provided by the Ministry of Health, which include counts from hospitals in all Governorates indicate that … [i]n June 2006 1,554 civilians died violently … The overwhelming majority of casualties were reported in Baghdad. In addition, the Medico-Legal Institute in Baghdad (MLI) [i.e. the Baghdad morgue] separately reported receiving 1,595 unidentified bodies in June. The total figure of civilians killed in Iraq, adding the figures provided by the Ministry of Health and the MLI, reaches … 3,149 in June 2006.
So June is clear. 3,149 in total = 1,554 per Ministry of Health + 1,595 per the Baghdad morgue.
For July, the NYTimes reported:
July appears to have been the deadliest month of the war for Iraqi civilians, according to figures from the Health Ministry and the Baghdad morgue … The total number of civilian deaths [in July was] 3,438 … The Baghdad morgue reported receiving 1,855 bodies in July, more than half of the total deaths recorded in the country. … The United Nations has been tracking civilian casualty figures by collating numbers from the Health Ministry and Baghdad morgue. …
So July is also clear. 3,438 in total = 1,583 per Ministry of Health + 1,855 per the Baghdad morgue.
Now read the USAToday report for August:
Preliminary figures show that violent deaths in Iraq dropped substantially in August from record levels the previous month, a Health Ministry official said Thursday. At least 973 violent deaths were recorded throughout Iraq as of Wednesday, Riad Abdul Amir of the ministry’s statistics bureau told the Associated Press. …
Still, the new figure represents a significant drop from a tally of 3,500 deaths in July reported by Deputy Health Minister Adel Muhsin. Muhsin said that was the highest monthly figure recorded since the war began in March 2003.
Violent deaths in the capital fell to 550 in August, from about 1,500 in Baghdad in July. That was the city’s lowest monthly tally this year.
Abdul Amir’s figures were compiled from reports by morgues and the Interior and Defense ministries.
Yes, I can read. The subsequent paragraphs which do say that the 973 is a combined number. I still think there is some confusion or error in reporting somewhere along the line. Possibly the phrase “by morgues” does not include the Baghdad morgue. Possibly Minister Muhsin is understating the numbers for political reasons. My best guess is that the 973 is comparable to the earlier month’s Ministry of Health numbers (of 1500+). And, note that even that interpretation would be a welcome one-third drop in these deaths.
The Iraq Coalition Casualties site reports a modest 7% drop in reported civilian casualties. The ICC only captures deaths reported in the media; it has consistently run about half of the UN total. I follow these trends fairly closely, and as the days of August went by (even on my vacation) I did not pick up any “dropping by two-thirds” pattern in the ICC data. So, is it possible that in August, for unknown reasons, the percentage of deaths that were reported by the media went from 50% of “total” to over 100% of “total.” I put “total” in quotes because, given the state of affairs in Iraq, the real number is not known, and by “total” I mean the UN numbers described above.
I know we’ve put several thousand more troops in Baghdad recently. But does anyone really think that the sectarian violence, of the nature we’ve read about in Iraq, just “stops” in one month, or suddenly drops by over two-thirds? Let’s try some theoretical math. 7,000 (approximately?) troops switched to Baghdad decreased violent deaths there by 1000. Can we deploy another 7,000 and end the violence altogether? … I admit to being “unserious” here. But my point is, does anyone find it credible that the re-deployment of several thousand troops has had such a marvelous impact, in less than thirty days? I find that hard to believe, but it could be so.
Presumably the UN will issue its report in the coming weeks and we will get a confirmation one way or another.
I welcome any suggestions or comments that might help clarify this.
Update: This from the Sept 4 NYTimes says:
And in Baghdad as a whole, the toll over the past week has been high, with the city morgue reporting more than 334 people killed or found dead from Aug. 24 to the end of the month.
Kinda hard to fit 334, from one week, from Baghdad’s city morgue into a country-wide total of 973. Very roughly, one week from the Baghdad morgue might be one-eighth of the total for the month (not one-third).
Declining Civilian Casualties?
Iraqi Civilian Casualties Down in 2005
An agonizing ‘gotcha’ moment
fact-checking Kos - Low grade civil wars
Freedom’s just another word for nothing left to blog