You Know More about Casualties in Iraq than the News Media Does

A reader pointed out that Kevin Drum was taking a look at casualty statistics from Iraq, and I noticed that Drum cited an article in the LA Times that reports civilian casualty counts from January through August. The article states that the information was acquired as follows:

The numbers are based on morgue, hospital and police records and come from officials in the ministries of Health, Defense and the Interior.

This offers another nice opportunity to check on the validity of the statistics I use based on media reports compiled by Iraq Coalition Casualty Count. I charted the numbers provided by the LA Times article to see what they suggest about recent trends in Iraq. Here they are, with the value for February upwardly adjusted to correct for the fact that it has only 28 days (which I always do for the ICCC data as well):


I suppose I should adjust the numbers for months with only 30 days as well, but the changes would be so slight that I usually don't. The dashed purple line represents the average number of civilian casualties from January through May (the 5 months preceding the onset of troop surge operations). The dashed blue line represents the average number of civilian casualties from June through August (the 3 months during which the troop surge has been operational). As you can see, civilian casualties are, on average, about 260 per month lower over the last 3 months than they were over the preceding 5 months.

The next chart shows the corresponding figures from Iraq Coalition Casualty Count:


It should be immediately obvious to you that, although the values differ (with those from ICCC being about 12% lower than the LA Times figures), the exact same story is told. According to the ICCC figures, civilian casualties are about 300 per month lower over the last 3 months than they were over the preceding 5 months. Here are the two charts shown side by side, in case you would like to compare them directly:


The only slight anomaly I noticed involves the April casualty figure from the LA Times article. I was immediately suspicious of it because it stands out from the other numbers. The other numbers look like the normal numbers you get when you count events in the real world, but the figure for April is different in that it comes in at exactly 1500 (not 1496 or 1503). That makes it look like a rough estimate, not a count. Still, it could be that it just happened to fall at that exact value even though it truly is a count. But that value also happens to be the only one that falls below the corresponding figure supplied by ICCC (which is 1521). That makes two strikes against the April estimate, so I think the true value is probably higher. But whether or not the April figure is accurate, the main point is that the LA Times casualty statistics and the ICCC casualty statistics (as well as the AP figures discussed here) tell the very same story. Given that the statistics were derived using different methods, it suggests (though it does not prove) that the results are valid.

The decrease in casualties over the last few months is not just a reflection of seasonal trends, as Kevin Drum implied here. This can be most easily appreciated by looking at the corresponding casualty chart from 2006:


As you can see, last year, casualties for June, July and August were, on average, higher than they were in the preceding 5 months (by more than 100 per month). At the time, I was hoping that this increase was just random noise in the data, not a genuine trend. Unfortunately, the results from later months showed that it was, indeed, a trend, one that continued right through the fall and into winter. My hope now is that the decreasing casualties of late are a sign of a trend as well. Only time will tell.

I know what you are thinking: alright already, the ICCC numbers are valid!! You are beating the "validity check" issue to death!!! Tell me something interesting!!!

OK, here's the thing: imagine that you had only the AP or the LA Times to go by (Kevin Drum appears to be in that boat). What can you learn from their casualty reports? Almost nothing. Yet, everything I know about the conflict in Iraq comes from my analysis of casualties. That's because, using the ICCC data, you can break down casualties by type and figure out how many are execution-style killings by Shiite militias in Baghdad, how many are suicide bomb attacks against Shiites throughout Iraq, how many are victims of mortar attacks, how many are killed by gunmen, and how many are collateral casualties from roadside bombs and other confrontations between insurgents and security forces. Looking at these details, you can sit at a computer half a world away from the action and figure out the difference between an insurgent and a terrorist. But all of this is meaningful only if the figures provided by ICCC are valid. And that explains my frequent attempts to validate their data.

Do you see my point? Given that the ICCC numbers are valid, now you can do more than any AP reporter or any LA times reporter can do. More than anyone at the New York Times or the Washington Post can do. They don't know that the ICCC numbers are valid, so they have to go find some anonymous Iraqi official to quote, and even then they only get the summary casualty statistics (not the critical details). But I seem to be the only person on the planet who realizes that the detailed information is out there, and you can go analyze those details for yourself (by the way, if you are going to try it yourself, read this post first!). It's strange to think that, with nothing more than a computer, you can be vastly more informed about todays news than entire news operations with vastly greater resources at their disposal. But when it comes to casualties in Iraq, that's how it is.

As I said, virtually everything I know about Iraq came from tracking casualties (and then thinking through the implications). If you don't track casualties in some detail -- and virtually no one else can because they don't know where to get the relevant information -- you'll be lost when you finally stumble across a few casualties figures in an article in the LA Times. Kevin Drum seems a little lost to me when he says:

Casualties are probably down a bit in Baghdad but up in the rest of the country. This jibes with the widespread prediction that one of the main effects of the surge would be to temporarily drive the insurgency (or some part of it, anyway) to areas of the country with lower troop concentrations. Put troops in Baghdad and the insurgents melt away and fight elsewhere.

No, it is not the insurgency that is moving away from Baghdad to other parts of Iraq. It is al Qaeda terrorists who are doing that. Terrorists who are not fighting a civil war, nor are they trying to restore the Baathists to power. They have something altogether different in mind. Moreover, civilian deaths were much reduced in August except for one thing: terrorist attacks against civilians carried by al Qaeda. Those deaths increased by quite a lot (as I showed yesterday), and that's why overall civilian casualties remained roughly constant from July to August. But to know that, you have to be able get at the casualty details. If you can't do that, you end up drawing silly conclusions, such as the idea that "the insurgency" is being driven away from Baghdad.

Early on in the war, right wing commentators made the mistake of referring to all of our enemies in Iraq as "terrorists." Left wing commentators made the mistake (and still make the mistake) of referring to all of our enemies as "insurgents." The fact that the Baathist insurgents and al Qaeda terrorists are both Sunni and were allied with each other until recently made it hard to distinguish between them (so people used one term or the other). By tracking casualties, I noted the difference between them long ago. It's time for everyone else to do that as well. That is, everyone should immediately cease referring to al Qaeda terrorists as insurgents. If you can't bear the thought that they might really be under the control of al Qaeda (because, in some vague way, that fact seems to help Bush), then at least refer to them as generic terrorists. After all, with their suicide bombers, they intentionally and indiscriminately slaughter innocent men, women and children. If that's not a terrorist, what is? If you think that a more accurate term for a killer like that is "insurgent," then you need help. Get some.

Despite my quibbles with him, I would like to credit Kevin Drum for taking note of a detail that I had overlooked with regard to U.S. military casualties in Iraq. I mentioned that military casualties were down in August, but they were down even more than I implied. He links to this article, which includes the following chart showing U.S. combat deaths (excluding deaths from accidents, which I had included):


As you can see, combat deaths were down sharply in August. Why? Presumably because the Sunni insurgents are increasingly working with American troops (instead of attacking them) as everyone finally takes aim at al Qaeda in Iraq. That is, "the insurgency" has not moved away from Baghdad; instead, Sunni insurgents and U.S. forces are fighting al Qaeda wherever they find them in Iraq.

It is good to be clear about the objective evidence in light of the report General Petraeus will soon deliver. Civilian deaths are down (not hugely, but undeniably, it would seem), military deaths are substantially down, sectarian killings in Baghdad are way down from peak levels, Sunni insurgents have started to cooperate with American troops as they have turned against al Qaeda, and indiscriminate killings of innocent civilians by al Qaeda are way up. That's where we stand today, and that's why Harry Reid is very wrong to suggest that "this war is lost." He knows that now, which is why he will deliberately downplay the military progress achieved against al Qaeda and focus instead on the lack of political progress. He will, in essence, be saying: "Iraqi politicans are not willing to reconcile with each other. Therefore, surrender to al Qaeda." That makes no sense to me, and I hope it seems equally nonsensical to the American public.