Pinpoint Inaccuracy
Kevin Drum and Dan Drezner puzzle over the massive increase in “collateral damage” in the last two decades. They find this puzzling in part because of, in Dan’s words, “a revolution in precision-guided munitions — which, in theory, was supposed to reduce the likelihood of collateral damage.”
Kevin’s theory is that some intentional killing of civilians simply gets reclassed as “collateral damage” because these days intentional killing of civilians is uncool. That seems a reasonable explanation of part of the phenomenon. For the rest, I think Dan and Kevin should consider the possibility that the “precision-guided munitions” story is mostly bullshit. The weapons aren’t necessarily as accurate as the press releases make them out to be. And even accurate weapons get fired by humans based on official and unstated doctrine. If doctrine holds that you’ll happily fire a bunch of missiles into a house with 20 civilians and one Official Bad Guy, then it’s very nice for you that the missiles are accurate, but not so nice for the 20 civilians. If you’re willing to fire one PGM into a car with one OBG and his family, and that car is on a busy road, again: you’re deciding to precisely guide a missile into a bunch of civilians, inescapably. This is the practice of the US and Israel, who are the chief employers of precision-guided munitions. If your actual small-unit practice in urban areas when coming under fire from unknown sources is “spray and pray,” then the fact that you also belong to a military with fancy warheads doesn’t keep nearby civilians alive. Piously declaring your intent to minimize civilian casualties while refusing to provide metrics on same – even claiming you don’t keep them – makes it impossible to manage toward lower civilian casualties.
Comment by Professor Coldheart —
November 19, 2009 @ 10:11 am
The weapons aren’t necessarily as accurate as the press releases make them out to be.
Eugene Jarecki’s Why We Fight touches on this point. The myth of the missile fired from a destroyer in the Persian Gulf and guided by satellites to smack Uday Hussein right in the moustache and completely bypass his neighbors is, oddly enough, a myth.
Comment by Anon —
November 19, 2009 @ 10:45 am
They may not be as accurate as the press releases make them out to be, but they are certainly much more accurate than before. I think a more likely reason is simply that PGM may get used much more than unguided weapons would have been used. Also, an accurate weapon doesn’t help much if your intelligence is inaccurate.
Comment by matthew h —
November 19, 2009 @ 11:15 am
Having crunched the numbers quite a bit on this, I think the correct answer is the one of yours I requote below.
Civilians mostly die “knowingly”, meaning that someone aims something that goes boom in their direction, knowing that they are there or very probably there, but decide that hitting it is worth it because of some valued target that may be there. I recall the countd of deaths of civilian Iraqis in the initial invasion were basically people in the area where command and control individuals from Saddam on down were believed located. (By the way, Lancet’s figures are bullshit and over the top on the whole Iraqi thing).
Exploding things or intense shooting at or among crowds of civilians is why civilian casualties get higher. (from atom bombs to artillery in market places to wedding parties hit by missiles to car bombs). Partly this is the result of anti-insurgent warfare, where insurgents disappear in a civilian area (and in alot of places kill civilians themseles, insurgents are a great part of the anti-civilian killing sometimes but especially in Iraq).
If you are having a hard time hitting the person you are aiming at, while you can indeed hit someone else, chances are youll hit nobody as civilians run and hide. But if you are blowing up the whole area, or laying down/spraying a huge amout of autmoatic fire, you’ll hit other people more easily. And generally one knows what the area is.
“Killed in the crossfire” is rare and not the main cause of large civilain casualties. Dispoportionate firepower targeted in a narrow area of civilians with wither full intent to kill civilians or knowingly accepting the risks of doing so, is the culprit. Negligence happens but it not the main problem.
This is it, here, and add suicide and car bombs on the guilt side of insurgents:
“If doctrine holds that you’ll happily fire a bunch of missiles into a house with 20 civilians and one Official Bad Guy, then it’s very nice for you that the missiles are accurate, but not so nice for the 20 civilians. If you’re willing to fire one PGM into a car with one OBG and his family, and that car is on a busy road, again: you’re deciding to precisely guide a missile into a bunch of civilians, inescapably.”
Comment by TGGP —
November 19, 2009 @ 11:22 am
I think the reason is that technological advancement has resulted in more long-distance weapons for us to use. I got into an argument about that here.
Comment by matthew h —
November 19, 2009 @ 11:24 am
Finally RTFA, I think their data needs more careful screening, also they make a fatal incorrect distinction that criminal lawyers shouldnt miss: they divide into intentional v unintentional. THats not how homicide works, its intentional versus KNOWING versus reckless/negligent.
Intentional means they want the civilians dead; thats rarer or hard to prove (though not for WWII bombing which most accept as intentional); KNOWING means what we are talking about here —
aiming with intent to not to kill the civilians but knowing (being “substantially certain”) that they are likely to die when the boom happens.
They dont want to kill Mrs OBL (or maybe they are happy to but it is not the mission objective) but they know she and the others are going to go. That’s alot of what appears to be happening.
Comment by matthew h —
November 19, 2009 @ 11:27 am
Wikipeida speaketh and these are the mens rea distinctions that need to be carefully done before simply dividing between intentional and unintentional, which is too crude, even for law (they add oblique intention too):
Direct intention – the actor has a clear foresight of the consequences of his actions, and desires those consequences to occur
Oblique intention – the result is a virtually certain consequence or a ‘virtual certainty’ of the defendant’s actions, and that the defendant appreciates that such was the case..[4][5][6]
Knowingly – the actor knows, or should know, that the results of his conduct are reasonably certain to occur
Recklessness – the actor foresees that particular consequences may occur and proceeds with the given conduct, not caring whether those consequences actually occur or not[7][8][9]
Criminal negligence – the actor did not actually foresee that the particular consequences would flow from his actions, but a reasonable person, in the same circumstances, would have foreseen those consequences
Comment by matthew h —
November 19, 2009 @ 11:29 am
Most modern civilian caualties are probably in the knowingly or recklessly sliding scale, but in teh simplistic unintentional versus intentional, people assume it is bettwen intentional and merely negligent (accidentally killed in the crossfire,etc.)
Comment by Glen Raphael —
November 19, 2009 @ 2:26 pm
TGGP comment brings up an interesting point – that precision-guided weaponry might be underwhelming wrt outcomes in a way similar to what happened with auto airbags and antilock brakes – we made cars safer, but that enabled people to drive faster and more carelessly, leading to an uncertain effect on traffic deaths. Thus: all else held constant, more precision might imply fewer civilian deaths, but the expected precision of the munitions is an input to the function that determines how much bombing we do. On the margin, if the bombs get better along any axis of possible improvement – cheaper, faster, longer range, more accurate – we should expect that would lead us to bomb more and in less certain circumstances than we otherwise would.
Comment by Glen Raphael —
November 19, 2009 @ 2:29 pm
Also: what has happened to population density between 1900 and now in war-affected areas? And are bad guys perhaps getting better at hiding in populated areas as a defensive and PR tactic?
Comment by Nancy Lebovitz —
November 19, 2009 @ 2:32 pm
I would think there’s more of an effort to count civilian deaths. I’ve never heard of anything else as careful as the Lancet study in Iraq.
That would be a contributing factor– all the other theories sound plausible, too.
Comment by Eric Martin —
November 19, 2009 @ 2:37 pm
By the way, Lancet’s figures are bullshit and over the top on the whole Iraqi thing
Correct me if I’m wrong, but the Lancet study did not measure civilian deaths, or collateral damage, as such, but rather excess deaths. I believe that is a very large distinction.
Without passing judgment on the methodologies and merits of the Lancet study.
Comment by Moe Blues —
November 19, 2009 @ 3:01 pm
On the margin, if the bombs get better along any axis of possible improvement – cheaper, faster, longer range, more accurate – we should expect that would lead us to bomb more and in less certain circumstances than we otherwise would.
In practice, it is just the opposite. We bomb MUCH less because we can accomplish the job with fewer bombs. The TanhWha Bridge in North Viet Nam was attacked dozens of times by hundreds of airplanes dropping nearly 1,000 bombs. All missed. However, when the Navy showed up with the first laser-guided bombs, the bridge was destroyed by two bombs in a single mission. It has been thus ever since: We no longer need to drop hundreds of “dumb” bombs on a single target to ensure destruction.
Are civilian deaths up or down for comparable conflicts? I say, pick any conflict that predates PGMs and compare it to Iraq I or II or Afghanistan. The differences are quite dramatic.
Comment by Keith —
November 19, 2009 @ 3:28 pm
(By the way, Lancet’s figures are bullshit and over the top on the whole Iraqi thing).
Oh good. I’m glad we’ve only killed 100,000 people for dubious and fabricated reasons instead of 1,000,000. That makes me feel much better.
Comment by almostinfamous —
November 19, 2009 @ 3:33 pm
like in washington DC, maybe?
Comment by Eric the .5b —
November 19, 2009 @ 3:42 pm
If the plausible death toll from the last several years were to drop to as low as 100K, that would be good.
Comment by KEn —
November 19, 2009 @ 4:15 pm
I agree with Glen Raphael and TGGP.
I’m guessing that there as long as “collateral damage” from each strike is ~20-50, it won’t make too much of a stir in the papers, so it’s no big deal.
If missiles get better, we’ll just drop more of them, or aim them into more populous areas. And the “collateral damage” will still be ~20-50.
Comment by Donald Johnson —
November 19, 2009 @ 6:43 pm
“By the way, Lancet’s figures are bullshit and over the top on the whole Iraqi thing).”
Yeah. You know this based on what, exactly? And which Lancet report? Lancet 1’s midrange figures are in good agreement with the study published in the New England Journal of Medicine–they both found about 100 violent deaths per day in the first 18 months, about three times higher than IBC’s numbers for that period.
To be clear, I’m not wedded to the 600,000 violent deaths of Lancet 2–I’m just surprised so many people on both sides of this debate are so certain of themselves. (Well, not really surprised–people often tend to be much more certain of themselves on such things than they should be.)
Comment by Donald Johnson —
November 19, 2009 @ 7:04 pm
I googled a bit and found this, Matt H–
link
I’ve read all those arguments before and refutations and refutations of the refutations–Tim Lambert’s blog was the place to go for all that. Nevermind. Agnosticism remains the right choice, IMO.
Comment by Donald Johnson —
November 19, 2009 @ 7:50 pm
Tim Lambert’s summary of the Iraqi death toll numbers from various studies. (I’m too lazy to reread, but I think he took the early studies and extrapolated them to the present in order to compare the earlier ones with the later ones, but I don’t remember how he did it).
link
Comment by Kolohe —
November 20, 2009 @ 7:21 am
The slowness of my internet connection is preventing me from downloading all the links to get context but this stat (in two different posts):
Skipping over the period of 1900-1990 is rather missing the big picture, to say the least.
Comment by Kolohe —
November 20, 2009 @ 7:40 am
Furthermore, this graph belongs in the bad graphs hall of shame. 1944 as an inflection point for civilian deaths? Really?*
You wouldn’t tolerate such sillyness from persons who didn’t share your viewpoints and preconceptions.
*yes, I know that the brackets are based on the data source. Making a trendline graph on that though, is Bad Stats 101.
Comment by ajay —
November 20, 2009 @ 8:00 am
I’d be rather dubious of a report of civilian casualties from many of the wars of the 19th century. Was the Qing Dynasty able to count how many peasants got killed in the Anglo-Chinese Wars? Did it even care?
Comment by LarryM —
November 20, 2009 @ 11:55 am
I think the following is implied by your post but not stated, and is perhaps the biggest factor: more accurate missles allows nations to rationalize their use in more circumstances. Without precision guided missles, and absent indescriminate bombing of civilians, which thankfully is mostly a thing of the past, targeting options are limited. But have (sort of) precision targeted weapons “allows” them to be used in more situations where “only” a (relatively) few civilians are at risk. Reprehensible to be sure, but I think a pretty accurate statement of the mindset of (at least) the U.S. and Israeli governments.
Comment by Mark Thompson —
November 20, 2009 @ 5:54 pm
I’ll add my voice to the chorus saying that more accurate weapons means more overused weapons. I prefer analogizing it to low-fat snack foods though. People start eating the low-fat snacks, then decide “hey, they’re low-fat, I can eat as much as I want.” Suddenly, they discover that despite their now low-fat diet, they can now serve Thanksgiving dinner on their gigantic butt.
Comment by Jackmormon —
November 21, 2009 @ 12:37 am
In the age of precision-guided bombs, do civilians still build and use air-raid shelters?
Comment by Frank —
November 21, 2009 @ 3:13 pm
My son served as an intelligence person in Afghanistan, callling in airstrikes based on reports from the field (this was before the recent increase in the use of drones).
He agonized over the possibility of civilian casualties and over whether or not the intelligence he got from the field was accurate.
Words fail me.
Comment by Glaivester —
November 22, 2009 @ 12:58 am
Another factor to consider is that the type of war we fight is much different now. We are fighting small groups of guerillas within a large civilian population rather than fighting traditional armies.
Moreover, there aren’t large traditional military targets either. It’s not like the insurgents in Iraq have a large base with a munitions factory next door that we can target.
Comment by The Angry Optimist —
November 22, 2009 @ 3:23 am
Is it just a coincidence that the cited portion of the study is from 1989 – 2009? Hmm…prior to 2001, what military operations of this size and scope were going on?
Right. None.
And lest we forget, the graphs at the link are a total and complete joke.
Comment by Jim Henley —
November 22, 2009 @ 8:38 am
TAO: All you’re doing is making clear the disproportionate responsibility of the US for those deaths. How do you imagine this as a win for you?
Comment by Barry —
November 23, 2009 @ 12:41 am
Glenn: “…we made cars safer, but that enabled people to drive faster and more carelessly, leading to an uncertain effect on traffic deaths.”
IIRC, the deaths per passenger mile driven has been dropping quite a bit over the past few decades. Disentangling the different effects of various changes might be difficult, but something’s certainly happening overall.
Comment by The Angry Optimist —
November 23, 2009 @ 3:18 pm
Yeah, if it went from “0″ collateral deaths in X year to “1″ collateral death in “X+1″ year, and that 1 was caused by the US, that would be disproportionate too.
you relied on some bad, sloppy math and bad, sloppy facts, including some personal cherrypicking on your end. Man up and admit it.
Comment by ajay —
November 23, 2009 @ 3:24 pm
Hmm…prior to 2001, what military operations of this size and scope were going on?
Right. None.
Wrong. For a start, you’ve got:
The Congolese civil wars, which by 2001 had killed hundreds of thousands of civilians;
and the Ethiopian-Eritrean border war in 1998-2000, which killed at least a hundred thousand combatants and civilians.
Comment by Eric Martin —
November 23, 2009 @ 4:00 pm
Does that mean TAO relied on some bad, sloppy history?
[Inquirng readers huddle in anticipation of the up manning]
Comment by ajay —
November 24, 2009 @ 6:49 am
On checking the original paper, it doesn’t include the Congo war, presumably because the author decided to class it as a civil war with international involvement rather than an interstate war. It does include Ethiopia-Eritrea, though.
Comment by whatevz —
November 24, 2009 @ 7:28 am
Casually tossing “Israel” into the mix is intellectually lazy at best.
Israel expends awesome amounts of resources and incurrs awesome risks to its own military and civilian populations to minimize the number of civilian casualties in its targetted strikes. The anti-Israel zeitgeist causes massive numbers of lies about Israel’s military to make their way into the mainstream press but the facts are that when Israel was targetting lone terrorist leaders they put awesome efforts into not causing harm to his family and when Israel was engaged in all-out 24/7 war they still went to greater lengths than any other nation in a state of total war to cause the least amount of damage to civilians, even when it was obvious (and has since been confirmed by non-partisan parties) that the enemy was attacking from civilian bases precisely on account of Israel’s moral decision to not fire back heavily into those areas.
For a nation that’s been actively threatened by some of the planets worst thug nations for the entirety of that nation’s brief existence, Israel has behaved more morally than would any other nation on Earth in a similar predicament. Whitewashing this stark fact out of intellectual laziness (or intellectual cowardice) isn’t worthy of you.
Comment by mb —
November 24, 2009 @ 1:22 pm
Translation: Henley’s an anti-Semite. Surprised it took so long for this accusation to appear.