Economics and similar, for the sleep-deprived

Monday, October 16, 2006

 
The Northern Ireland Numeraire

More on the Lancet study �

Everyone knows that the standard unit of geography is "an area the size of Wales" (except for military purposes, where for some reason it is "an area the size of France"). Back in the dim days of history when the first Lancet survey came out, Marc Mulholland wrote something which got me to thinking that perhaps a similar standard unit for civil wars ought to be "the worst year in Northern Ireland". By the way, I'm conscious that we're talking about real deaths of people who had families and loved ones here. There is, frankly no way of writing this post which isn't as sick and tasteless as hell, so I'm just basically going to go for it. All I can do is beg the forgiveness of my Northern Irish readers and hope that, at least, the fact that their civil war appears to be at an end makes it a bit more bearable. Sorry in particular to Marc, who I must emphasise doesn't bear any responsibility at all for the use I've made of his numbers.

In the worst year in Northern Ireland, 1972, 496 people were killed in acts of violence related to the Troubles. If we scale that up for the difference in population between Iraq and Northern Ireland (27m/1.5m = 18x) and for the time covered by the Lancet study (39 months/12m = 3.25x), then we can say that, if the situation in Iraq since the invasion had been exactly as bad as the worst year in Northern Ireland, then there would have been 29,016 deaths by violence. Almost certainly coincidentally, this is the lowball number that George Bush gave when initially trying to rubbish the Lancet study. It does not, to be honest, seem very credible.

The IBC count of 43k-48k, would correspond to the worst twelve-month period in Northern Ireland, immediately following on from the introduction of direct rule (which would in turn correspond to the average death rate attributable to Israeli security forces in Gaza between 2000 and 2005). To me, at least, this kind of points out how low this count really is. The figure of 65k which has been thrown around quite a lot and appears to be rather tenuously sourced from Iraqi government data, would correspond to Iraq since the invasion being about as bad as the worst single month in Northern Ireland, July 1972 when nearly a hundred people were killed.

So basically, all these lower figures would be consistent with the country of Iraq having been, on average, about as lawless and dangerous as the really bad times in Northern Ireland. I don't think that this is remotely plausible. July 1972 was a horrible time from all I can read (I wasn't born at the time), but even then, it seems that there was at least a semblance of normal life going on in a sense that appears to be utterly absent from Iraq. I'd also note that British troop numbers scaled up from 17,000 to 29,000 during that period; to maintain the same troops/population ratio in Iraq, there would need to be just under 400,000 coalition troops in Iraq, about three times as many as there actually are.

On the other hand, the central Lancet projection would suggest that on average over the three and a quarter year period covered, in terms of violent deaths alone, Iraq has been roughly twenty times as bad as 1972 in Northern Ireland. Specifically, it was roughly ten times as bad for the period covered by the first Lancet study and is currently 36 times as bad. Is this credible? To be frank, I don't know; I think that these numeraires break down when you have to multiply them by too many times. But looking at these numbers makes me very confident indeed that the low numbers are completely incredible; they're not describing anything like the state of civil war that Iraq appears to be in.

The bottom line here is that, as I say, the Iraq war isn't just a terrible mess, like British policy in Northern Ireland during the Troubles. It's a massive disaster, well on its way to becoming a disaster of world-historic proportions like Vietnam. The big and frightening headline numbers are just the logical numeric expression of that fact; if someone isn't proposing to argue with me on the fact that it's a total disaster then they can think what they like about the numbers.

Update: Jamaica had 1476 murders in 2005, out of a population of 2.7m. Scaled up in the same way, this gives a number almost exactly equal to the Iraq Body Count "high" estimate. Safe European Home = Rock the Casbah?
0 comments this item posted by the management 10/16/2006 03:25:00 AM