Abu Muqawama: Post
Abu Muqawama retains its autonomy and the views and beliefs expressed within the blog do not reflect those of CNAS. Abu Muqawama retains the right to delete comments that include words that incite violence; are predatory, hateful, or intended to intimidate or harass; or degrade people on the basis of gender, race, class, ethnicity, national origin, religion, sexual orientation, or disability. In summary, don't be a jerk.
On the Casualties of War (Updated)
John Tirman has an important if flawed op-ed in today's New York Times. He urges U.S. military and political leaders -- as well as the general public -- to be honest about civilian casualties in war. Tirman argues that U.S. military officers need to be wary of civilian casualties for strategic reasons, and here the two of us are in violent agreement. Tirman also argues that the U.S. public and its leaders need to consider the total human cost associated with war for moral reasons, and here too we are in violent agreement. Whenever I speak about the war in Iraq -- whether it is over dinner with friends last night or on NPR a few weeks ago -- I always make sure I mention the terrible loss of Iraqi lives. We Americans have to be honest about this. Last night, someone asked me if I thought the Iraq War had been worth it, and though I said the Iraq war had accomplished certain things (the fall of Saddam, a nascent democratic system of government), it most certainly had not been worth it. The three pieces of data I went on to cite were a) the $1 trillion spent, b) the 4,484 U.S. military lives lost, and c) the tens of thousands of Iraqi civilian lives lost. I could have gone on to cite coalition casualties, the Iraqi refugee crisis, and wounded soldiers and civilians, but you get my drift: I am sympathetic to the aim of Tirman's op-ed.
But then Tirman writes this:
In 2006, two separate household surveys, by the Iraqi Ministry of Health and by researchers from Johns Hopkins University, found between 400,000 and 650,000 “excess deaths” in Iraq as a result of the war. At the time, however, the commanding general in Iraq put the number at 50,000 and President Bush had claimed in late 2005 that it was just 30,000.
As Tirman has to know, that Johns Hopkins / Lancet survey was incredibly controversial when it was released and remains controversial today. It relied on cluster sampling, in Iraq, at the height of that country's civil war. I cannot think of a poorer environment in which one could do that kind of survey. Yes, it was peer-reviewed, but an academically sophisticated methodology cannot compensate for poor data. (Garbage in = garbage out.) Both Gen. Casey and Pres. Bush were likely much closer to the mark, as the iCasualties figures from the very height of the war in Iraq -- 2005-2007 -- are way lower than the figures from either of the studies Tirman cites. (And if Tirman thinks the Iraqi Min. of Health had the capacity, in 2006, to accurately measure the cost of the war on the Iraqi civilian populace, he needs to spend more time in peacetime bureaucracies in the Arabic-speaking world. I apologize for painting with such a broad brush, but those with experience dealing with large state bureaucracies in Egypt or Syria know of what I speak.)
Tirman's op-ed is basically a call for the United States to use violence more selectively, and it's a pity he overstates his case (as tends to happen in New York Times op-eds), because I agree with him. As has been demonstrated time and again, the use of indiscriminate violence in civil war environments confuses the population, scrambles incentive structures for behavior, and tends to inflame the population against the force using the violence. Selective violence is much more effective.
That's the strategic argument. The moral argument is that the U.S. public needs to understand the total human costs associated with its wars. That may lead the United States to be more selective as to when it applies U.S. military power abroad and how it does so. On the other hand, it might also lead the United States to think carefully about how it ends its wars as well. There is a fashionable sign in my neighborhood, for example, that reads "End the War in Afghanistan." I assume this sign is meant to read "End U.S. Involvement in the War in Afghanistan," because I myself am unsure as to whether or not the U.S. withdrawal will ameliorate or worsen the conflict there. Progressives like Tirman should keep that in mind: the U.S. military is only one actor in environments like Iraq and Afghanistan, and the U.S. presence is not the only driver of conflict. It is even possible -- whisper it -- that increased U.S. combat presence and operations might actually serve the interests of the civilian population in some cases. That's certainly the case, at least, in most stabilization operations.
Anyway, my congratulations to John Tirman for this important op-ed.
UPDATED: One of the folks in the comments section points out that Tirman directed the funding for the Lancet/JHU study. Well, that explains it! (I wish he would have disclosed this small but significant point in his op-ed.) Tirman apparently believes between 800,000 and 1.3 million Iraqis were killed in the war, which is a simply incredible claim. No one else puts the number that high. The Associated Press (110,600), the Iraq Body Count Project (103,536 — 113,125), and the Wikileaks logs (109,032) all put the number much, much lower. At what point does someone admit that their numbers just might be off and that their own study had deep flaws? I mean, only 87,000 death certificates were issued in the worst years of the war (2005-2008). Tirman might be the only guy left who references the Lancet/JHU study as having been sound.
Academics, journalists and
Academics, journalists and liberals (I repeat myself) will keep citing the Lancet Lie until it's accepted as fact, because it serves the narrative.
people died. that's enough to
people died. that's enough to know.
killed as a result of a useless american invasion, can be added to that.
the heck with academic quibbles about methodology.
crikey .. building a career?
What is the point? There is
What is the point?
There is no argument about civilian injuries. I do not think that any US army has gone into any conflict with the sole intent of doing damage to civilians. It is not a secret that civilians get damaged. Use to be that was part of war, now we have cashiers following the US army around to pay for damages. Where is the responsibility for the people that bred these nut cases that ran jets in the Trade Towers. If that never happened, Mohammad's opium shack would have never been fragged.
Seems to me that the discussion should be focused on a strategy to completely avoid the damage, like not doing Libyan NFZ's. It really should be about national interest. This 9/11 business has really gotten a life of its own.
America is fighting a small faction, why are we wasting so much resource on this? This is no longer about George W. Bush. If these wars are an investment what is the return to the US taxpayer? Getting to pay for more for the same is not an answer it is an insult.
There really needs to be an American discussion of what we are putting in and what is coming out. Nation building, as a reason, is just not good enough. Humanitarian claims are and excuse. Denial of area to AQ is a poor reason for the treasure lost and to be spent. AQ is an idea, not a place.
This is getting super power stupid.
How stupid can it get? Libya, they ask the US to do the damage and do some damage themselves so all the people in America can pay to chase their problems.
Why am I paying for this?
Andrew, FYI, Tirman is
Andrew,
FYI, Tirman is *certainly* aware of the JHU/Lancet controversy: he was one of the principal investigators on the 2006 Lancet study! To this day he contends -- in both academic and popular writings -- that his 650,000 number is right.
Given the importance he gives to that study in the op ed, I think was not not just an oversight but unprofessional to not alert readers that he was one of the authors of that study given that it's so at odds with other independent attempts to estimate Iraq casualties, not just at odds with US military claims.
Andrew, I agree with all your
Andrew, I agree with all your points. But here's the thing: It won't wash in the streets. Too nuanced. Too many difficult things to balance.
We need to recognize that while the war can do some good things, the cost in Iraq was too high, but not as high as this guy's saying, but too high nonetheless, though if we weren't in Afghanistan exacting those costs, the cost might even be higher, so keep that in mind, too.
All quite right. But when they're ginning up a war with talk of mushroom clouds and existential threats and whatever tomorrow's aluminum tubes and yellow-cake may be, the guy who follows Andy Card's advice and "talks to America like it's a ten year-old" is gonna win every time.
So, should we simply despair? A little, yes: "The heart of the wise is in the house of mourning . . ."
But maybe we should also listen to this: "Much protest is naive; it expects quick, visible improvement and despairs and gives up when such improvement does not come. Protesters who hold out for longer have perhaps understood that success is not the proper goal. If protest depended on success, there would be little protest of any durability or significance. History simply affords too little evidence that anyone’s individual protest is of any use. Protest that endures, I think, is moved by a hope far more modest than that of public success: namely, the hope of preserving qualities in one’s own heart and spirit that would be destroyed by acquiescence." Wendell Berry
And who knows? Maybe something good can come from that.
Aren't the news-based studies
Aren't the news-based studies and the Lancet study reporting different things? The news based studies give a conservative -lower limit to the number of violent deaths- (only those dead which make it into the news, or international news, are counted), while the Lancet study tried to give an -estimated total of all excess deaths-. The news based studies should understate the deaths several times, because newspapers are neither interested in nor capable of recording all violent deaths; the question is whether a factor of ten is reasonable. I haven't read any other serious attempt at an -estimated total- although that Wikipedia article links to two or three.
My inexpert guess is that if 100,000 violent deaths have reached the news, then probably several hundred thousand more Iraqis have died since 2003 than would have died under Saddam.
Your comments, in the fullness of personal knowledge, about the failings of Arab bureaucracies seem to support that large numbers of deaths have not been recorded, or if recorded have not made their way to Baghdad.
So, what are you going to do
So, what are you going to do about it Andrew?
“Expectations Management” and
“Expectations Management” and “Perspective” are perhaps in short supply today. While I would in no way encourage unnecessary suffering or death, if one looks at how wars were fought over the last 300 years and compare the estimates with today’s conflicts I would submit that on the whole we are causing far less collateral damage than in the past. That is a very laudable thing, expecting perfection is a dangerous road to travel that will ultimately weaken all parties to the delusion.
Visitor says, "if one looks
Visitor says, "if one looks at how wars were fought over the last 300 years and compare the estimates with today’s conflicts I would submit that on the whole we are causing far less collateral damage than in the past."
Certainly true. But not very relevant. Our communications technology, our valuation of human life, our humanitarian sensibilities, our ability to mobilize political movements around scandalous events, and the susceptibility of even great and powerful nations to the effects of those mobilizations, have all been revolutionized in the last 300 years, too. It's just not the same game.
And besides. If the government blew up your house during a family reunion because they were trying to kill some terrorist, would you just shrug and say, "Well, at least they're not Genghis Khan?"
Charlie Ford We have evolved,
Charlie Ford
We have evolved, certainly true however if our “enemies” have not then to what extent does our advancement ensure our ability to reduce casualties (on both sides) to levels that are socially acceptable to us? If at the end of the day we have advanced to an extent that we are no longer willing to do what is needed to actually win, whatever form that may take, then…
“If the government blew up your house during a family reunion because they were trying to kill some terrorist”. After having spent over 27 months in Afghanistan and Iraq I am quite confident that I would end up being one of the guys that the government found itself needing to hunt down. Even with that much time in theater I still improve my understanding of that AO constantly. Back to your statement though, that is why when Genghis Khan/Stalin/etc went into the “conquering other places” business they often killed off entire families and or villages at one time. If you kill the entire village people still get the message however they are somewhat less likely to try to come after you. I am in no way advocating reverting to slaughter as a method of pacification regardless of its proven effectiveness. What I do hope for is that those who advocate for perfection develop some course of action that will work when applied by US Army troops interacting with relatively uneducated persons in a place where they do not speak the language. The US Militaries rotational cycle virtually ensures that troops never go to the same place twice and due to Force Protection Concerns they are restricted from engaging with hardly any local persons except the cleaners on base…
Actually, the only argument
Actually, the only argument for saying the war was worth it was that it made it possible to lift the sanctions in place on Saddam, wich had up til then killed a serious amount of civilian innocents through lack of acess to medicines etc. But except for that, Phase IV? What a clusterf**k.
When it comes to doctrine, I do hope the memory of Fallujah does include the fact that Total Force Engagement in urban enviroments is not a very smart choice tacticaly as well as being a very brutalizing method of fighting. Tripoli has hopefully taught us that a siege is more efective in the case of urban conflict...?
Fnord "made it possible to
Fnord
"made it possible to lift the sanctions in place on Saddam"... again to what end. If we left too soon for political reasons then our efforts are far more in question based on a long term view of the issue. I think we can all agree that perhaps things should have been done different. Commanders will always want more capability. More capability provides them the ability to accomplish the mission in a more nuanced manner. If we went into Iraq with 400,000 troops’ things would have gone differently, in what ways and to what extent we will never know.
Urban engagements, especially in ethnically diverse areas, being accomplished by outsiders both culturally and linguistically, will always be messy at best and a disaster at worst. If you do it do go big or go home. As to Libya I will let you know in 5 years how that turns out. Right now there are a significant number of outlying issues to include the numerous MANPADS and other toys that got loose in part due to no one being there to secure them. Libya was kept whole by a semi-brutal guy…sound familiar to Saddam and for largely the same reasons. I would not be shocked if that country fragments in some significant manner. It may be ok if the go with a solution similar to the “3 states in 1” that was discussed with Iraq, short of that I would look to that area not being close to stable for some time.
Visitor, I hear what you're
Visitor, I hear what you're saying, and these are real--and difficult--questions.
Take this, eg: "If at the end of the day we have advanced to an extent that we are no longer willing to do what is needed to actually win, whatever form that may take, then…"
What is winning? If you're in a firefight, you know what it is. But what about from the perspective of the US as a whole? Winning is more ambiguous then: we usually want, at the end of the day, to eliminate more enemies than we create, ensure as best we can we leave a peaceful, just, economically viable nation that will be on good terms with us in the future, and so forth.
If we leave a nation embittered, looking for revenge, or merely wrecked, and a disaster for its neighbors and region, that, someday down the road, costs us more in blood or treasure than we would have lost doing nothing in the first place, have we "won"?
Arguably, eg, we have far more productive relations with the socialist Republic of Vietnam today precisely because we lost, than we would have had we persisted in that war till doomsday. Now, we have an ally (of sorts) and a trading partner. then, we spent blood and treasure. Should we have invaded the North and pummeled them into submission? Would that have resulted in a net gain, over what has resulted?
"If you kill the entire village people still get the message however they are somewhat less likely to try to come after you." Except now we have a global village. Even if you kill every Iraqi, you'll have the Indonesians coming at you, you'll rile up the Indian Muslims, etc., etc.
"What I do hope for is that those who advocate for perfection develop some course of action that will work when applied by US Army troops interacting with relatively uneducated persons in a place where they do not speak the language. The US Militaries rotational cycle virtually ensures that troops never go to the same place twice and due to Force Protection Concerns they are restricted from engaging with hardly any local persons except the cleaners on base."
Yes. We should really think twice--three times--about whether we can do these wars right, and if we can't, maybe try something else.
I encourage anyone here to
I encourage anyone here to listen to the 'This American Life' story on the JHU/Lancet study.
http://www.thisamericanlife.org/radio-archives/episode/300/whats-in-a-nu...
I always dismissed it as crackpot but that segment forced me to re-examine it.
Just a couple of
Just a couple of clarifications. I was not a PI on the second Hopkins study; I commissioned it but had nothing to do with conducting it or writing the article that appeared in The Lancet. I do believe it has more rigor than other estimates, but of course the sampling could have been unintentionally mistaken. The fact that there were two household surveys at the same time and they found 400,000-650,000 excess deaths, and that there was a lot of killing and dying to come, suggests very high numbers for the entire war. The low estimates of ~100,000 are based on news accounts in English and after action reports, both of which rely on crude and very partial detection techniques, and one never knows what's being missed. Holding these up as standards is quite foolish. This is detailed in my book. We need to do much better. It's important to know how violence affects the war and the society, and only scientifically sound surveillance, done frequently and independently, will do that.
Yeah, Tirman's claims about
Yeah, Tirman's claims about this are slippery and misleading in several ways. First, note the manipulative descriptions of the two surveys he cites. One is "by the Iraq Ministry of Health" while the other is by "Johns Hopkins University". The first survey is actually by a group of organizations called the Iraq Family Health Study Group, one of which is the Iraqi Health Ministry, but the designers, supervisors, trainers, and lead authors of the study were from the World Health Organization. The only thing the Iraqi Health Ministry did was provide a few hundred doctors to go to the selected houses and collect the data. Yet for some reason the study is described based on its data collectors. Meanwhile the Lancet study is described based on the parent organization or affiliation of the academics who designed, supervised etc. Why the double standard here? Well, since the WHO study concluded that the Lancet study wildly overestimated the number of violent deaths, and calling it WHO sounds more prestigious, best to call it something else. And hey, maybe the "Iraq Health Ministry" had an agenda and so forth (wink, wink).
The data collectors for the Lancet study were a mysterious doctor named Riyad Lafta that nobody in the media or anywhere else is ever allowed to talk to, and a group of "Iraqi data collectors who have asked not to be identified". So if we use the same naming standard here based on who collected the data, rather than who designed, supervised, etc., the "Johns Hopkins University" study should probably be called something like The Anonymous Iraqi Data Collector survey. But somehow that standard doesn't apply here.
Also, the WHO survey did not "find" 400,000 excess deaths. In fact, it didn't estimate excess deaths at all and no such estimate has ever been openly described or defended in any academic context such as a peer-reviewed journal. It has been reported in a media article that one of the WHO authors once offered a figure just under this in some off-the-record conference, but it isn't clear in what context or what methods or how reliable he actually thinks it is. And one author saying something at a conference is not the same as an estimate being reviewed and published in an academic journal. I have my doubts that such an estimate could even be published in a journal such as NEJM, because it is so speculative. The survey itself does not support this number. If you use the rates found, even with missed clusters added, you don't get anywhere near 400,000. You have to do big upward adjustments to the survey data to get there. And what of the statements in the NEJM report on the survey such as " The underreporting of deaths was expected to be lower for more recent years.", and in the WHO Q&A, "reporting of deaths longer ago is less complete". If these statements in their reports are true, then the counterfactual comparison of pre-war to post-war rates does not actually result in a number of "excess deaths". It results in a number that may be partly excess deaths and partly anomalies from a changing level or reporting bias across the recall period, where it is unclear how much of the number is one or the other. That is, if the statements of the authors about this are true, doing the counterfactual leads to an inflated estimate of excess deaths, since their data on deaths longer ago (i.e., pre-war) are "less complete" than the more recent data (post-war).
But Tirman ignores all this and just asserts that the study "found" this number that it never published and which has these noted problems. Then he just presents this as a range with the Lancet's number, so 400,000-655,000. But again this is slippery and manipulative. The difference between these two estimates is not just the difference between these two abstract numbers. Numbers have no meaning unless you know what the numbers represent. In this case, the Lancet number of 655,000 consists almost entirely of violent deaths (601,000), the vast majority adult men, killed by bullets and bombs and such. But this WHO number of 400,000 would have to consist primarily of non-violent deaths from disease or other causes that kill people in war-time and peace-time, with only 150,000 coming from violent deaths, and the vast majority would not be adult men, but a more general population demographic. The two numbers aren't measuring the same people. They are measuring two completely different Iraq's. And it shows that at least one of these studies has to be completely wrong. If one is at all credible, the other is simply not. Tirman sneakily elides this gigantic problem by presenting only two abstract numbers, and implying that the difference is simply a numerical difference between the two, ignoring causes of death and demographics entirely, as if the estimated war victims were some kind of fungible currency, and it doesn't matter that the people represented by the two numbers are two completely different sets of individuals.
The next manipulative thing Tirman does is use slippery pronouns to quietly move goal posts around. So he gives two excess deaths figures and then says, "At the time, however, the commanding general in Iraq put the number at 50,000 and President Bush had claimed in late 2005 that it was just 30,000." So it's "the number" and "it". But this "it" is not "excess deaths". Their "it" is not Tirman's "it". In neither instance were they talking about excess deaths. They were talking about the number of people killed in the violence.
Then in his reply above he's manipulative regarding counting sources. "The low estimates of ~100,000 are based on news accounts in English and after action reports, both of which rely on crude and very partial detection techniques, and one never knows what's being missed." He mostly seems to be talking about Iraq Body Count here, but this presents a too limited account than is actually the case. And it is also, as Abu notes in the update above, death certificates issued that come to similar numbers. Note too that while Abu says that 87,000 certs for violent deaths between 2005 and 2008, the Lancet study claims about death certs would require that about 550,000 were issued for violent deaths between 2003-2006. It doesn't come close to adding up.
Anyway, he then goes into some real double-think. On one hand "one never knows what's being missed," while on the other Tirman somehow knows that it's "very partial". Regardless of the double-think, the claim that "one never knows what's being missed" isn't really true, except in the sense that you can almost never have certainty with these kind of complex questions. IBC has done analysis regarding what is likely missed and these were largely born out by the Wikileaks material. That is, what is missed tends to be smaller scale incidents with one or two deaths each, while it is very unlikely that things like mass-casualty bombings are missed. This and other types of analysis can provide a pretty good idea of what is missed. But Tirman's complaint is ultimately strictly true for any approach, including his household surveys. The complaint boils down to that you can't know for sure how close a count like IBC is getting to the true number. Well, that's also true of the surveys. They can give an estimate that seems credible, but you can't know for sure how close (or how far) they got to the true number. In the case of the two surveys Tirman mentioned for example, one of the two got the violent death estimate very badly wrong, and far beyond the margins of sampling error. So things can go badly wrong with these surveys, and it's not always easy to tell what really went wrong, or if something has gone wrong or not. So ultimately it would be just as true to say with Tirman's surveys that you can't know for sure how close (or how far) they got to the true number. Yet this only bothers him with regard to the counts.
Ultimately, I think it's hard to own up to the fact that you've wasted everyone's time and money on a lemon. So Tirman won't budge. Unfortunately, it becomes even worse when he smears other researchers over this, as he recently did on radio when calling researchers from Boston University's Costs of War project "morally reprehensible" because they didn't buy his lemon.
SM wrote: "Aren't the
SM wrote:
"Aren't the news-based studies and the Lancet study reporting different things? The news based studies give a conservative -lower limit to the number of violent deaths- (only those dead which make it into the news, or international news, are counted), while the Lancet study tried to give an -estimated total of all excess deaths-. "
In this case, the answer is not really. "Excess deaths" is different in that the number can include both violent deaths and non-violent deaths above a pre-war baseline. But in the case of the Lancet survey it did estimate violent deaths specifically, not just excess deaths, and the excess death estimate wound up being almost entirely violent deaths anyway. So in that case they aren't measuring different things so much.
However, the most prominent news-based study is Iraq Body Count, but in that case another difference is that IBC is for violent deaths of civilians, while the survey estimates of violent deaths are for all violent deaths including combatants. So obviously the survey estimate should come to a larger number since it's including more types of violent deaths, but this difference couldn't come close to bridging the gap with the Lancet number.
On the other hand, if you consider the "400,000 excess deaths" estimate that Tirman imputes to the WHO study - which is one excess death you could draw from it, but you could just as credibly, or even more credibly draw other lower ones from it imo - that estimate winds up being mostly non-violent deaths. So in that case the excess death estimate would be measuring something very different than a source like IBC.
"The news based studies should understate the deaths several times, because newspapers are neither interested in nor capable of recording all violent deaths; the question is whether a factor of ten is reasonable."
Many news sources are actually very interested in recording and reporting all violent deaths. Some sources report lists of all they can gather about it every day. How capable they are to do so is another matter, and would tend to vary. I don't think it follows however that combined news reporting should "understate the deaths several times" at all. The IBC approach is pretty broad. News sources are listed here: http://www.iraqbodycount.org/analysis/reference/sources/, and there are a lot. They also supplement news reports with "NGO-based reports, along with official records that have been released into the public sphere." So this can include stuff like Human Rights Watch reports or UN reports, and official figures from Iraqi hospitals or morgues. And then now they are including the US military records from Wikileaks and so forth. This is quite broader than what you might presume when thinking about "newspapers". It may well still miss a substantial number, but it does not seem at all self evident or even likely that it should "understate the deaths several times", imo.
Ces site web informatif!
Ces site web informatif! Grand merci! Merci pour un bon moment la visite www.cnas.org. C'est vraiment un plaisir compréhension d'un site web comme celui-ci regorge d'informations sympa. Merci!
Add your comment