Davies on six studies of deaths in Iraq

Many thanks to Tim for bringing these sorts of articles to our attention.

I realize that this is somewhat off-point, but I am trying to gather data on pre-war rates of death under Saddam. The Lancet Study provides one such estimate but only for a limited window. I often see claims that "millions" were killed by Saddam, but I would be curious to see citations for these claims. Roberts et al, alas, were not doing surveys in Iraq in 1991.

In doing this, I am obviously wrestling with the concept of "excess" deaths. In the context of the Lancet study, this is defined as the difference in death rates in the year or so before and after the war. The problem with this approach is that Saddam might have been on his best behavior in the run up to war. If his average killing rate in, say, the 1990s was much higher than in 2003, the 2003 estimate might not be the best to use.

Again, I am just looking for empirical facts and claims. How many people did Saddam kill over what time periods?

I apologize for (trying to) hijack this thread, but there are few better locations for discussion of this topic.

David Kane:

The problem with this approach is that Saddam might have been on his best behavior in the run up to war.

I have problems with that argument, interesting though it is. From what I remember, Saddam did try to convince the international community (some what too late) that he did not have WMD or means to deploy them (i.e. dismantling rockets), but there was no need for him to cease his killing. The plight of the Iraq people under his rule was not the reason/justification given for the invasion prior to the event. Only after the failure to find WMD in Iraq was the plight of the Iraq people used as a justification for the war by the US/UK governments.

I recommend to check also this column by the math corrector, John Allen Paulos, on bird flu virulence and Iraqi civilian deaths.

David Kane, the Davis et al study on Costs of War in Iraq (easily goggled) cite deaths under Saddam as 500,000 so they claim the invasions death rate per year is about comparable. Their claim not mine.

Opponents of the war (like TL) always cite the largest figures while Lancet gave confidence limits on their estimates from under 10,000 to 194,000. Supporters of the war go for much lower figures.

Despite TL's protestations to the contrary I don't think we have a clue what the approximate number of deaths is. The Lancet student is a rough sample survey. It doesn't pretend to be accurate though it is often discussed as if it is.

The difficulty is that in times of war or during oppressive regimes people have incentives to conceal deaths. Unfortunately thiss does not mean the highest estimates offered are accurate.

Harry Clarke wrote:
> Davis et al study on Costs of War in Iraq

You're talking about the Davis, Murphy, and Topel paper. It's interesting that you cite it -- one of the experts they cite to back up their estimate is Richard Garfield, described as "the most credible source for estimates of how sanctions have affected mortality among Iraqi children." Garfield is one of the "et al." in Roberts et al.

"I often see claims that "millions" were killed by Saddam, but I would be curious to see citations for these claims."

Total civilian and military casualties in the Iran-iraq War, which Saddam was primarily responsible for starting were between 1 and 2 million.

The US State Department has said Saddam was responsible for 300,000 Iraqi civilian deaths. This includes approximately 100,000 Kurds killed in the Anfal ethnic cleansing campaign and 80,000 shia killed in the post-Gulf War uprising in southern Iraq.

Both the US government and independent human rights grouspestimated in the lead-up to the 2003 invasion that the Iraqi government was responsible for 2-3,000 extra-legal deaths per year.

In another piece of his last updated on 2 March, Davies says that coalition forces have killed anywhere between 70,000 and 500,000 people in Iraq since the war began. Including 30,000 to 275,000 children.

Those figures are so far out of line with any others that one is bound to question Davies' credibility.

It appears that Davies is a polemicist rather than a statistician. If you look at the titles of the stuff he writes, and where he publishes them, it is not hard to see which side of the argument he is coming from.

http://www.worldproutassembly.org/archives/2006/03/burying_the_lan.html

I agree, Davies offers no justification for these figures. He just hits you with:

'Allowing for 16 months of the air war and other deaths since the completion of the survey, we have to estimate that somewhere between 185,000 and 700,000 people have died as a direct result of the war. Coalition forces have killed anywhere from 70,000 to 500,000 of them, including 30,000 to 275,000 children under the age of 15'.

These estimates won't become credible until they are corroborated by being repeated in at least 3 left-wing blogs. BTW my range of estimated deaths from the US invasion of Iraq (including deaths the US was responsible for but didn't cause directly) now goes from 7,800-700,000.

The figures are important because they are vital for establishing whether Iraq is better off with invasion than it would have been with continued slaughter by Saddam at 10,000-30,000 deaths per year. My guess is that it is about line ball at around 40 deaths/day comparing but that the death rate has escalated markedly over the last 6 months.

Whatever the statistics the outcome is obviously horrible for the unfortunate people of Iraq. They suffered terribly under Saddam, the US got rid of him and they are still suffering.

Yes they are suffering dreadfully. And the prime cause of their daily suffering is the continual attacks by unreconstituted Ba'athists and Islamist terrorists. Plus a good deal of help from "friendly neighbours".

harry

How do the wild assertions that Davies made become credible simply by being repeated in 3 left-wing blogs ? That is just an example of a Lie being half-way round the world before Truth has got its boots on.

There are some interesting comments on this article about the way the Lancet figures have been misused :

http://www.logictimes.com/civilian.htm

dumbcisco, you know you haven'yt hiit the nail on the head when you have to explain your (grim) humour. I was being ironic - repetition is the way factoids in this universe become fact.

The post you give is just a twist on the low figures given in Iraq Body Count.

harry

Sorry. Your ironic description of how BS becomes "fact" is spot on.

Harry Clarke, how come when Richard Garfield produces an estimate of deaths as a result of the war you dismiss it, but when Davis and co take Garfield's estimate of deaths as a result of the sanctions and make some assumptions that Garfield probably would not agree with and then come with a 10,000-30,000 per year figure you accept it?

Can we take it as agreed that the Davis upper figure of 500,000 is absurd ?

Tim Lambert, It's a guess - a hypothetical, a counterfactual - what deaths would have been had Saddam remained in power. Saddam had incentives to conceal his killings.

BTW accepting this guess (even if I did) would have nothing to do with accepting death estimates post-invasion which seems to be what you are implying.

It is discouraging to me (and I hope its wrong) that I get the feeling that estimated casualty rates are being used to evaluate prople's attitudes to the invasion - if you go for really high guesses you are very anti-invasion etc. If you say you don't know you are being identified as a war-monger or, as I was recently described. as a "cheerleader for mass murder".

Ian Gould wrote:
"Both the US government and independent human rights grouspestimated in the lead-up to the 2003 invasion that the Iraqi government was responsible for 2-3,000 extra-legal deaths per year."

I have not seen these supposed claims from human rights groups. The vast majority of killings under Saddam were heavily concentrated in the 1980s (incidentally, while he was receiving diplomatic, financial and military support from the United States and the UK for his invasions and slaughters). After roughly 1991 (after that support had ceased), the rates dropped off significantly, and nothing on the same scale had been seen since, with the lowest points being the years directly preceeding the invasion.

Amnesty was using figures like "scores" (dozens) and "hundreds" to refer to people killed by the Iraqi regime in the years directly preceeding the war. And these figures included official death penalty cases (not just "extra-legal"), as AI still considers the death penalty a human rights violation. In short, I've seen no indication from human rights groups of figures as high as even 2-3,000 in Iraq for many years.

Harry Clarke wrote:
"The figures are important because they are vital for establishing whether Iraq is better off with invasion than it would have been with continued slaughter by Saddam at 10,000-30,000 deaths per year."

Again there is no basis for this claim other than constructing an extremely misleading "average" sweeping deaths from the 1980s up to 2003, as if the Iran/Iraq war were still ongoing, which is presumably how the claim is derived. According to Human Rights Watch there hadn't been mass killing going on in Iraq for many years, and as noted above, AI was using figures like "hundreds" and "scores" in its yearly reports for 2002, 2001, 2000..etc.

In short the "continued slaughter by Saddam at 10-30,000 per year" that supposedly would have been "continuing" if not for the invasion is a fabrication, a sleight of hand that bears no resemblance to reality. Such a rate would not be a "continuance" of anything. It would be a remarkable change in the conditions in Iraq before the war, amounting to an explosion in death rates, out of the blue, amounting to numbers not seen for over a decade in Iraq.

And we're supposed to be thankful that the US invaded and produced that explosion, thereby preventing the imaginary one from "continuing".

Good editing is key to better presentation! - I'm taking the liberty of editing Harry Clarke's comment to read "They suffered terribly under Saddam (aided and abetted by the US until the first Gulf War), then twelve years after the war the US tried to murder but only succeeded in getting rid of him (and killing an awful lot of innocent Iraqis as collateral damage) - and they are still suffering". Well, I think it reads beautifully now!

Harry, you insinuated that I was only supporting the Lancet estimate because I was "anti-war" and now you claim that it "is discouraging to me (and I hope its wrong) that I get the feeling that estimated casualty rates are being used to evaluate prople's attitudes to the invasion".

Please explain.

Tim, If you take that as an 'insinuation' I withdraw it. I have no basis for making that claim in relation to you. I notice that opponents of the war go for large estimates and supporters for lower estimates but have no evidence that it is this that drives your figures.

JoshD how could the 10-30,000 figure be anything but a 'fabrication'? It is a counterfactual, a hypothetical. Davis et al. are trying to compare deaths with the US invasion with an estimate of the deaths that would have occurred had the US not intervened and Saddam remained in power. They are trying to evaluate the cost of intervention with the cost of continued 'Containment' which they see as the plausible alternative. Unless you have the ability to re-run history how else are you to proceed?

You might reject the methodology (I don't but, like you, I find the empirics difficult) but that wasn't the issue.

"JoshD how could the 10-30,000 figure be anything but a 'fabrication'? It is a counterfactual, a hypothetical."

It's a fabrication even for 2002, 2001, 2000, 1999...etc. which is not hypothetical. We know the figure is wildly unrepresentative of the years directly preceeding the war. Nothing hypothetical about it. Yet we're supposed to believe that rate would "continue" without the invasion, even though it wasn't "continuing" before the invasion.

Unless we assume a sudden and dramatic explosion in deaths in 2003 without the invasion, a return to 1980s level killings which hadn't been seen in over a decade, the "hypothetical" is quite implausible. The claim is just absurd for the reasons I described.

Furthermore, the time periods they must be including to arrive at their bogus "hypothetical" 10-30,000 average can hardly be based on "containment". "Containment" only began after the first Gulf War, while the average must be smuggling in all the deaths from the 1980s, while Saddam was receiving diplomatic, economic and military support from the West for invasions, wmd development and internal slaughters. That's hardly "containment".

Actually, it appears your source says:

"How does this tally of human misery compare to war? Under the policy of containment in effect since the end of the Gulf War, premature Iraqi deaths have numbered at least 10,000 per year and probably two or three times as many."

So there's the 10-30,000 figure. But it's speaking of "premature Iraqi deaths" during "containment", which means we're primarily talking about sanctions and the resulting deprivation, not killings by the regime as you'd implied, and as I was therefore assuming.

In this case we still have a problem of disparate time periods as the early/mid 90's were much worse for this than the late 90's and onward, after the installation of the oft-maligned Oil For Food program, which did help ease the suffering and deprivation for the Iraqi people.

But the other problem with this logic is of course, that it was only the US insisting on maintaining the sanctions by the end. Britain was sort of going along but didn't really want them, and all the others stopped supporting the policy long ago. If the US wanted to stop those supposed 10-30,000 yearly sanction deaths, they could have made a phone call and done so.

And lastly, since we now know the figure is referring to "premature death" and not to killings by the regime, there's no indication that health conditions have improved except possibly for a few areas in the country, and many indications that it's gotten worse in most of the country. So the net gain for even this 10-30,000 would have to be a big 0, and most likely with a considerable loss added on top.

Everyone really should read this piece. I found it here:
http://gsbwww.uchicago.edu/fac/steven.davis/research/War%20in%20Iraq%20…

It's actually quite astonishing how wrong they were about just about everything they speculated about the war. First, throughout they simply assume Saddam had stockpiles of WMD (the purported justification for continuing the sanctions that the authors are using to sell the "lives will be saved" argument), which we know was false.

Then there's this howler:
"The resulting estimate for the cost of containment is $380 billion. This dwarfs any reasonable estimate of U.S. war costs."

Keep in mind too that the estimate for "containment" was over a projected 33 year period. The war has probably exceeded that in just 3.

Josh D, Cool it my friend. The counterfactual was maintaining sanctions. This was judged by Davis et al as the most plausible alternative counterfactual.

The deaths under the counterfactual were estimated to apply post invasion. This is hypothetical not data.

The study you cite is 2003. Their updated study is 2006 - it is much longer piece and yes is different. They changed many things. Perhaps inconsistently? I have not carefully compared.

The information for the study they use is what THEY CLAIM could reasonably be expected to be known at the time of invasion. To say that WMDs were not in Iraq and to criticise US intervention because they are not there is fallacious if the expectation that they were there was reasonable in 2003 - and they, NOT ME, claim it was reasonable to suppose they existed in 2003. Otherwise THEY CLAIM one is exercising ex post wisdom by being smart after the event. Thus it is an ex ante study as of 2003 based on information they see as available then.

All costs under invasion and containment were extrapolated into the future over the same horizon and discounted at 2%. The $380 billion cost of containment compares with, from memory, their cost estimate of invasion, of about double that. THEY FIND the two sets of costs to be comparable by accounting for failures of the containment policy and for reduced US security risks. They value the 9/11 attack and value something like a dirty nuclear attack at twice that. They then claim getting rid of Saddam reduces the probability of such an attack by 4% giving present value savings of $80 billion. Oh yes, and before you jump down my throat this procedure has been widely criticised - at the Becker-Posner blog, for example,

The estimated invasion costs reasonably comparable across a range of studies - Bilmmes/Stiglitz say $1-2 trillion which is a trillion more than Davis et al but half of that is based on a $5-$10 increase in oil prices as a consequence of the war and on higher US interest rates - they attribute these as a cost of invasion. Davis don't. They are mostly around $1 trillion if you leave out macroeconomic costs.

All this is in the 2006 study - its on the web. BTW I apologise for capitalising several bits - I know its bad web manners but I am describing their study which I have put a lot of effort into understanding. To repeat - the approach seems to be the best I've seen in terms of approach but, like you, I disagree (or find questionable) many empirical assumptions. It could be argued, for example, that invading Iraq reduced US security by limiting options elsewhere at later times. Again is this ex post wisdom based on nuclear strike issues in Iran? This possible 'loss of options' cost is not included in the Davis et al. study although it is mentioned in Bilmes/Stiglitz.

Cheers

Whilst a dictatorial regime may have reason to hide the excess deaths, the US run Coalition of the Willing seems also to want to hide the excess deaths, or at least make no serious effort to keep track of them. It is in the absence of a clear body count that this debate is occurring. Clearly good information makes for better decisions and neither overestimates nor underestimates help in evaluating the effectiveness or otherwise of intervention policies - decisions about future adventures in intervention could well rely on such knowledge, with anywhere from a few thousand (as if thousands of human lives should ever be considered "few") to hundreds of thousand or more dependent on the outcome of such decisions. Actually, given that it's only deaths being estimated, clearly far more living people are adversely affected than just those killed, so let's just say millions of lives are dependent on such decisions.

Fine, I'll check out the 2006 version, though i don't hold very high hopes. Anyway, the original point of my jumping in was your claims that:

"The figures are important because they are vital for establishing whether Iraq is better off with invasion than it would have been with continued slaughter by Saddam at 10,000-30,000 deaths per year. My guess is that it is about line ball at around 40 deaths/day comparing but that the death rate has escalated markedly over the last 6 months."

We now know that the 10-30k figure is based primarily on an averaging of primarily "premature" sanction deaths over the "containment" period, not really "continued slaughter by Saddam" which was relatively minimal during that period.

As far as that is concerend, your "vital" question should be answered. There's no indication that the conditions leading to "premature death" during containment/sanctions have improved. There's evidence (such as the Lancet study on this point) that these health trends have only worsened still further since the invasion (except perhaps in some areas of the North).

So there is no evidence of any gain whatsoever in terms of the supposed 10-30k "containment" average. In fact, there's evidence that it's higher now. So we can just add the explosion of violent deaths right on top of that supposed 10-30k. And it wouldn't even matter which figure you choose for that. You'd get the same answer to your "important" and "vital" question. In terms of loss of life, your "guess" appears to be about as absolutely wrong as one could be.

*Sigh*. I love reading the apologists of US-UK aggression, attempting desperately to put a good spin on the carnage inflicted by more than 50,000 coalition bombs while - as Krusty put it succinctly - sending all western complicity during Saddam's period of worst atrocities (1980-90) down the memory hole.

We can debate until kingdom come the actual extent of US-UK genocide in Iraq - and how the west views its victims as 'unpeople' when they conflict with the economic agenda of the rich and powerful. If 9-11 was indeed a 'bloodbath' and an 'atrocity' (which I fully agree that it was on both counts), then how would we describe the killing of 30,000 Iraqis, or 60,000, or, as it seems most likely - 100,000 plus? An accident? An aberration? An anomoly? An exception to the rule'? When the 'rule' includes such horrors as US-inflicted death in Viet Nam, Cambodia, Korea, Nicaragua, Chile, Panama, Brazil, Iran and support for brutal regimes in Nigeria, Indonesia, Colombia etc., then I can only shake my head in disbelief at some of the dumbass comments here, making all sorts of excuses for just the latest example of western mass murder.

US-inflicted deaths in Korea ????? Sheesh, please learn some history.

Dumbisco,

Earth to Dumbcisco. Wakey, wakey. Perhaps you need a little refreshing up on the Korean war. How many civilians died under US bombardment, DB? Estimates range up to 2 million; some say even more. There's no doubt that, as in Iraq and Viet Nam, the civilan infrastructure was deliberately targeted.

Most importantly, DB, what constitutes 'industrial state killing' in your lexicon? How many Iraqi deaths from coalition airstrikes and the ground assault (ignoring the after-effects of the assault on the surviving civilian population) constitutes a slaughter? A bloodbath? I asked this above; why not hazard a guess? Several NGO's in 2002 estimated that an attack on an already devastated Iraq would put at risk the lives of up to half a million people. Did the Bush regime care? They didn't give a damn - the Iraqis join a long list of peoples around the world whose lives were meaningless because they stood in the way of imperial power and expansion.

What is clear from Tim's weblog is that there is a definite attempt by some like DB to downplay yet another example of carnage iflicted by western (mostly US) forces. The issue is constantly defined as those who were 'for' and those who were 'against' the war. But this is the wrong way to frame it. The issue should be framed around those who support preventive aggression by a super powerful state against an effectvely defenseless state in clear violation of international law and the resulting mass murder, or those who oppose this. The US, following on proudly from previous imperial powers (ancient Rome and Nazi Germany come to mind) claims that they needed to attack Iraq in 'self-defense'. This was precisely the same hollow rhetoric used by Rome as it plundered neighbouring lands and Nazi Germany when it aggressed against Poland and Czechoslovakia (read the Nuremberg transcripts and you'll see this was their lne of defense at the trials).

If DB wants to learn a little bit about the recent history of the Middle East, I'd suggest Robert Fisk's fabulous "The Great War for Civilization" - its a long tome (almost 1300 text pages) but details the western role in the tragedy that has enveloped the region over the past 200 years.

I expect this will go down like a lead balloon here :

http://www.qando.net/details.aspx?Entry=3652

Jeff Harvey

You recommend Robert Fisk. Now isn't that surprising ?

And you also cite the alarmist NGO predictions re. Iraqof 2002 that proved groundless. That is my point - on matters of statistics, people will tend to quote the figures that support their political view. Your political views veer to a particular extreme, and it is unsurprising that any statistcs you cite do the same.

If you cast the defence of South Korea as some sort of imperialism by the US, I'd like a few pints of what you are drinking.

Dumbcisco, I happen to think that the US defense of South Korea was, in the long run, a worthwhile act. And it was defensible in just war terms, since North Korea invaded. But Jeff Harvey is correct about the massive number of civilian deaths caused by US bombing--you can find this in mainstream histories if you look hard, though nobody knows if the true figure was several hundred thousand or over 2 million. Curtis LeMay (quoted in one of Bruce Cumings's books) once said his planes killed over a million civilians.

It's also a fact that South Korea in the 50's was every bit as brutal as North Korea. Both sides massacred civilians (and so did US troops, quite apart from the bombing). And the war didn't really start in 1950--both sides were killing (both across the border and internally, I gather) in the years leading up to the invasion.

South Korea is now a much happier place than North Korea, but that's because the South Koreans overcame their own dictatorship. The US role in that country has been mixed--we protected them from the Stalinist North, but we also killed massive numbers of civilians and propped up a fascist regime for decades. It's quite possible for all sides in a war to be guilty of massive war crimes. Happens a lot, in fact. As in Iraq. As for statistics there, I think order-of-magnitude estimates are about as accurate as we are going to ever have, and that's not unusual in wars. (As in Korea.)

If we didn't have an actual Seixon/dumbcisco-type commenter, we'd have to make one up.

Go git 'em, Jeff.

Best,

D

My father left Australia in 1915 to fight on the Western Front. I think he would be amazed at some of the anti-West views current in Australia these days, which transcend zaniness. No names, no packdrill, but the degree of closed-mind anti-US spleen is saddening.

"There are some interesting comments on this article about the way the Lancet figures have been misused :
http://www.logictimes.com/civilian.htm"

Yeah, that's the guy who claims < 1,000 civilian casualties, counting only civilians directly killed by US troops. Interesting, for sure.

"There are some interesting comments on this article about the way the Lancet figures have been misused :
http://www.logictimes.com/civilian.htm"

Yeah, that's the guy who claims

Interesting, my post got truncated at the less than sign, presumably it's expecting a piece of HTML to follow.

Anyway, that's the guy who claims fewer than 1,000 casualties for the war, if you consider a casualty to be a civilian directly killed by US troops. I don't know if he'd count a civilian killed by a building falling on him as a result of a US bomb that didn't actually hit the victim itself, though.

'The Iraq Living Conditions Survey

This survey was conducted by the Ministry of Planning and Development Cooperation of the Coalition Provisional Authority in April and May 2004 and was published in May 2005 by the U.N. Development Program. The "UNDP" imprimatur and the large sample size gave credence to its reassuringly low figure of about 24,000 "war deaths." Tony Blair's biographer John Rentoul told me that, in his opinion, this survey was definitive. However, its estimate of war-deaths was derived from a single question posed to families in the course of a 90-minute interview on living conditions, and the Norwegian designer of the survey has said that this number was certainly an underestimate. More than half of the deaths reported were in the southern region of Iraq, suggesting that it captured deaths in the initial invasion rather than in the violence that followed. In any case, after the invasion itself, the period covered by this survey was one of relative calm, and the two years of increasing violence that have followed are unaccounted for.'

That is considered an argument? Does that even need criticising? Surely it is sufficient in itself to discredit the author, anyone who agrees with it, and anyone who ever mentions the Lancet study as anything other than a footnote in the history of studies opf propaganda and media manipulation?

As has often been pointed out, it is not incompatible with tha Lancet study. This is because the Lancet study contains virtually no information, and so is compatible with just about anything short of the US invasion causing corpses to rise from the ground or the entire population being killed twice.

This means if you put together the UN study and the Lancet study, the result is essentially the same as considering the UN study alone. So anyone who is aware of the UN study, but uses the Lancet one instead, should strongly be suspected of attempting to sell a bill of goods by dishonest means.

Dumbcisco:

but the degree of closed-mind anti-US spleen is saddening.

If you want something other than criticism of the current US actions: The UK must have killed thousands in the bombing of Hamburg & Dresden.

dumbisco
The Great War was Great as in large, not great as in "what a great way to celebrate western civllisation". Still, at least they did a reasonable job of keeping it away from civillians while turning young men into chum.

I thank several people for posting useful links. The best one that I found led to this, an article by Stephen Cass (with a Ph.D. in Iraqi history from Oxford). He wrote:

Along with other human rights organizations, The Documental Centre for Human Rights in Iraq has compiled documentation on over 600,000 civilian executions in Iraq. Human Rights Watch reports that in one operation alone, the Anfal, Saddam killed 100,000 Kurdish Iraqis. Another 500,000 are estimated to have died in Saddam's needless war with Iran. Coldly taken as a daily average for the 24 years of Saddam's reign, these numbers give us a horrifying picture of between 70 and 125 civilian deaths per day for every one of Saddam's 8,000-odd days in power.

Again, my purpose here is just to get the best empirical estimate. Note that this estimate does not, I think, include Iraqi (or Iranian) soldiers. Questions:

1) Is 100 deaths per day a fair estimate of Iraqi civilian executions during the whole of Saddam's reign? Again, if you have a better estimate, please provide a link.

2) Is there good date on a time trend available (as several commentators mentioned above)?

Thanks to all who responded.

No, 100 executions per day is not a reasonable estimate. Amnesty International says the number per year was "scores" in the years leading up to the invasion.

but Tim, it might (or might not) be a reasonable estimate for the 1980s. And surely you couldn't fail to support a 2003 invasion that might prevent deaths from 1988!

But, I'd be curious to see how this "documentation" of 600,000 supposed "civilian executions" would fare under the kind of examination given to current Iraq figures. Cass appears to be the only person to cite it, other than by repetition in right wing blogs citing Cass.

The organization cited for the "documentation" seems to have had a website at some point (www.iraqcenter.org), but it appears to be offline. So no data to examine, no corroboration or citations from any mainstream human rights organizations afaik. Just an Iraqi group claims to have documented 600,000 executions, at least according to Cass.

For the folks for whom that evidence is sufficient to make a "reliable estimate", then surely the 2003 study by the Iraqi group People's Kifah, which documented a total of 37,137 civilian deaths just for the first 3 months of the war (mid-March to mid-June), is an equally reliable estimate for the invasion:
http://www.wanniski.com/showarticle.asp?articleid=2855

That leads to a total of about 400 deaths per day, or 4 times the per day rate of the apparently equally "reliable estimate" for Saddam's time.

Or maybe this is just nonsense and we don't really have any way to say how reliable either of these figures are.

DB does what all apologists of US agression do when confronted with incontrovertible evidence of western mass murder. He/she resorts to the old smear that the critics are "anti-American". I suppose DB reckons that the millions of American born critics of the lawlessness of Bush and his precedecessors are also "anti-American". Most of the prominent academic critics, indeed, are. I am sure that they are joined by many of the American bloggers here - people like Dano - who are proud of their country but also believe that there has to be accountability.

To repeat for the umpteenth time, the "Iraq war" wasn't. It was the same old story of a reckless, imperialistic superpower exporting death and carnage to a crippled, effectvely defenseless state in support of an economic agenda that mostly benefits the privileged few. That's why the Lancet debate misses the point. There are those, like DB, who feebly try to defend this genocidal foreign policy which has safely found a niche under the 'war on terror'. The fact is that the US-UK war party couldn't care less about human rights and democracy. This tired old rhetoric is constantly rehashed by the well-oiled state-corporate media apparatus.

If you want to understand what US-UK foreign policy is all about don't listen to the semantic lies of Bush, Blair, and the industrial killing machines they command. Read what the nation state/government planners have to say. These are the beaurocrats who are determining global policy decision making. And what you'll find if you read these documents - at least those that are declassified and readily available - is scarcely if ever a concern for human rights and democracy. The documents I've read - and many more that historians like Mark Curtis have read and written about - always express concern that nations whose resources we and our multinational corporations covet will 'embrace nationalist forces' that attempt to 'utilize their resources and the profits that derive from them for the benefit of their own populations'. This will 'conflict with the interests fo western businesses and corporations'. Thus, instead of supporting in other nation states the internal distribution of national wealth to benefit all sectors of society, our governments are instead concerned about it! Why? Because these forces will conflict with the interests of our corporate elites! And these documents make it clear that western powers should do everything in their power to 'maintain a strong political influence' over the governments of these countries. In document after document, the same theme recurs. Thus, the debate over the Lancet estimate of US-UK mass murder in Iraq is largely irrelevant, because it was always considered by the governments of the coalition to be a 'price worth paying' in order to gain control of an area rich in energy resources. But of course there was mass murder: whether the final tally is 60,000 or 100,000 Iraqi civilians or more, Bush and Blair and their minions are complicit in a massive international crime.

Priceless:

"The war was the hard part. The hard part was putting together a coalition, getting 300,000 troops over there and all their equipment and winning. And it gets easier. I mean, setting up a democracy is hard, but it is not as hard as winning a war."
(Fox News Channel's Fred Barnes, 4/10/03)

"What's he going to talk about a year from now, the fact that the war went too well and it's over? I mean, don't these things sort of lose their--Isn't there a fresh date on some of these debate points?"
(MSNBC's Chris Matthews, speaking about Howard Dean--4/9/03)

"Over the next couple of weeks when we find the chemical weapons this guy was amassing, the fact that this war was attacked by the left and so the right was so vindicated, I think, really means that the left is going to have to hang its head for three or four more years."
(Fox News Channel's Dick Morris, 4/9/03)

and many, many more. http://www.fair.org/index.php?page=2842

"E. was sitting at the other end of the living room, taking apart a radio he later wouldn't be able to put back together. I called him over with the words, "Come here and read this- I'm sure I misunderstood..." He stood in front of the television and watched the words about corpses and Americans and puppets scroll by and when the news item I was watching for appeared, I jumped up and pointed. E. and I read it in silence and E. looked as confused as I was feeling."
"The line said:
Ùزارة اÙدÙاع تدع٠اÙÙÙاطÙÙ٠اÙ٠عد٠اÙاÙصÙاع ÙاÙاÙر دÙرÙات اÙجÙØ´ ÙاÙشرطة اÙÙÙÙÙØ© اذا Ù٠تÙ٠برÙÙØ© ÙÙات اÙتØاÙ٠اÙعاÙÙØ© Ù٠تÙ٠اÙÙÙØ·ÙØ©
"The translation:
"The Ministry of Defense requests that civilians do not comply with the orders of the army or police on nightly patrols unless they are accompanied by coalition forces working in that area."
"That's how messed up the country is at this point."
http://riverbendblog.blogspot.com/

Of course 600,000 deaths are worse than 100,000. Unless of course one of those 100,000 happens to be your six year old daughter.

David Kane,

d-squared has blogged about the question you raise, but not recently. Most likely that means he has no new data, but if you e-mail him or lob a question into a comment thread on his blog I'm sure he will respond.

Soru, your own argument seemed too fuzzy to make you a plausible judge of other people's credibility. The point was that the UN survey asked only one question about war deaths and so we don't know for sure what the respondents had in mind. I might not count a murder if asked that question--I might only think of people I knew were killed by bombs or artillery or by foreign troops or car bomb. Since we know (first from the Lancet study and also from news stories from Iraqi morgues) that huge numbers of people are being and have been murdered by unknown assailants (criminals, death squads, insurgents, Americans?), a vague question about war deaths might leave out a lot.

As for the Lancet study, it sounds like you're just repeating the argument about its wide confidence interval and how that supposedly makes the study worthless. I doubt anyone here cares enough to go through all that again.

Er...Soru what the hell are you doing arguing on a blog about statistics, a subject on which you are immaculately ignorant? You make the extraordinary claim that 'the Lancet study contains virtually no information'. Yes, because obviously peer reviewed journal articles frequently contain 'no information'.

To quote from the article:

'To estimate the relative risk, we assumed a log-linear regression in which every cluster was allowed to have a separate baseline rate of mortality that was increased by a cluster specific
relative risk after the war.We estimated the
average relative rate with a conditional maximum
likelihood method that conditions on the total number
of events over the pre-war and post-war periods, the
sufficient statistic for the baseline rate. We accounted
for the variation in relative rates by allowing for overdispersion
in the regression. As a check, we also used
bootstrapping to obtain a non-parametric confidence
interval under the assumption that the clusters were
exchangeable.'

OK for starters: explain to me the difference between 'parametric' and 'non-parametric'. Then tell me what 'regression' means and 'log-linear'. Then go onto Confidence Intervals.

Then take us through your understanding of the statistical analysis of the data. The, please, tell us where in your opinion the authors went wrong, and then please give us your own ideas as to how a further study could deal with the 'problems' of the Lancet study.

This sort of shoot from the hip dismissal might work with the innumerate drones at Harry's Place, but some of us look for that strange little thing called 'evidence' to back up seemingly wild and implausible claims (and believe me it don't get much more implausible than the claim that the Lancet study contains no information).

I ask, "Is 100 deaths per day a fair estimate of Iraqi civilian executions during the whole of Saddam's reign?" Tim writes:

No, 100 executions per day is not a reasonable estimate. Amnesty International says the number per year was "scores" in the years leading up to the invasion.

1) It does not matter, to my question, whether the rate was high or low recently. (It may matter to other issues, but I am not concerned with this factual question.)

2) If 100 is not a reasonable estimate, than what is? (It is fine to say that no estimate is unreasonable, that the uncertainty is too great. But Tim seems to imply that he knows that 100 per day is to high. How does he know this?)

3) I appreciate Kevin's link to Daniel Davies, who is clearly a genius. Davies seems comfortable with an estimate of 35 per day or so. Would this be a reasonable estimate? Tim's citation of "scores" per year suggests that he might disagree with Davies. True?

Thanks again to all for their help. I am sad to report that the Lancet authors (and the editors associated with the article) do not seem that interested in helping me (or anyone) to replicate their work. But I will press on regardless.

"1) It does not matter, to my question, whether the rate was high or low recently. (It may matter to other issues, but I am not concerned with this factual question.)"

Then you aren't concerned with the question of how the war has affected iraqi lives. If the death toll was high in 1988, what is the relevance to the numbers killed since the 2003 war?

David: for 1979 to 2003, it seems that 35 per day is as good a figure as you're going to get: 200,000 Kurds and 100,000 Shiites killed in the course of crushing uprisings between 1988 and 1991; add say 5% for people murdered in the normal course of rule by terror and there's your 35 per day. But that's horribly sketchy and I'm doubtful that you will find anything to compare it with for the postwar period. Rule has been decentralised to militias and we don't have much information about their activities.

Since Roberts & Co. were doing a mortality study, not a study of violence per se, I'm wondering what you're getting at? Not only is the subject of study different, but the methodology isn't comparable at all.

joshd: Of course I am concerned about how the war affected the lives or Iraqis. But, in this post, I am just trying to focus on one empirical question.

Kevin: What I am getting at, indirectly, is the issue of "excessive" mortality as the term is used by the Lancet authors. We need to have a sense of the what, in the counterfactual world in which invasion did not occur, the death rate would have been. This is, obviously, a much harder issue than the (still hard) question of how many Iraqis were killed during Saddam's rule.

Again, at this stage, I am not claiming that, because 35 died per day during Saddam's rule, that rate would have continued in the counterfactual world of no invasion. As Tim and many others point out, most of those deaths occurred a decade or more ago. At the same time, it does not strike me as unreasonable to speculate that the best counterfactual estimate is not, as the Lancet implies, the 18 months before the war (when Saddam was behaving himself) but a similar period of Western back-down, say post Gulf War I.

I am not arguing that 100 or 35 or 0 is the best estimate. I just want to get a sense of the range of reasonable values. Knowing how many people Saddam killed helps in that effort.

might not count a murder if asked that question--I might only think of people I knew were killed by bombs or artillery or by foreign troops or car bomb.<./I>

Well, precisely. When people ask how many people died in a war, they want an answer to that question, not the question you think should have been asked instead.

The correct and mathematically correct answer to the qestion 'what is a million squared?' is 10^^12, but subsituting that mathematically correct answer for the one asked is a lie, and would not be less of a lie if the calculation were 12 pages long and subject to detailed disputation.

To make the same point another way, try running the Lancet analysis across the handover of power to Mandela in South Africa. I think you will find plenty of Broederbond members who will agree with your position.

Murder rates go up when a dictatorship is ended, either you accept that as a consequence of the change, or you don't. Most South Africans, and most Iraqis, do, and I am not sure anyone else has any moral right to override that decision.

So, then, Soru, there were maybe 24,000 people killed by wartime violence, not counting the many thousands more killed by increased gang violence, all in the first year and therefore before the summer bombing of Fallujah and the fall conquest of that town and all the violence since then. This is consistent with the Lancet numbers, as has been pointed out before.

I think the majority of Iraqis welcomed the overthrow of Saddam and were probably willing to swallow the initial several thousands of deaths to accomplish that. But now the death rate may be comparable to those of Saddam's worst years, for all we know and from what I've read they're not happy about that. I'm not saying that the Iraqis as a whole would agree with the antiwar movement in the West, because I'm not sure they would, but they probably would welcome a bit more soul-searching on our part regarding how we've gone about "liberating" them. And maybe something a bit harsher than soul-searching--war crimes trials, truth commissions, that sort of thing.

Incidentally, Soru, as you might notice, my opinions aren't quite what you apparently thought they were regarding the Iraqis. But don't let that stop you from tossing out those pointless irrelevant insults.

I thought I'd come back and concede one thing to Soru--I was antiwar, in spite of what I suspected many Iraqis felt in March 2003. But it seemed like better-than-even odds that any "liberating" we'd do would blow up in their faces.

"Murder rates go up when a dictatorship is ended, either you accept that as a consequence of the change, or you don't. Most South Africans, and most Iraqis, do, and I am not sure anyone else has any moral right to override that decision."

Did I miss the part where a majority of Iraqis requested the US to come in and topple Saddam, and damn the anarchy and chaos, they wouldn't mind if their kids got killed as long as it was in the service of a democratically elected fundamentalist Shiite theocracy? I remember the part where the largely Kurdish resistance requested we aid them and when we said we would they got gassed, but that was a different war.

Gee, I wonder if this could be relevant?

"IRAQ: UN report cites vast under-nutrition among children
08 May 2006 14:08:32 GMT
Source: IRIN
BAGHDAD, 8 May (IRIN) - One in three Iraqi children is malnourished and underweight, according to a report released by the United Nations Children's Fund (UNICEF) in Amman on 2 May.
"Under-nutrition should not be accepted in a country like Iraq, with its wealth of resources," said UNICEF Special Representative for Iraq Roger Wright from the Jordanian capital, Amman. Wright added that ongoing insecurity served to deter parents from visiting health centres for essential services, while many health workers had been kidnapped or killed in different parts of the country.
According to the report, a full 25 percent of Iraqi children between six months and five years old suffer from either acute or chronic malnutrition. A 2004 Living Conditions Survey indicated a decrease in mortality rates among children under five years old since 1999. However, the results of a September 2005 Food Security and Vulnerability Analysis - commissioned by Iraq's Central Organisation for Statistics and Information Technology, the World Food Programme and UNICEF - showed worsening conditions since the April 2003 US-led invasion of the country.