Northern Ireland Numeraire - update

So, how are things in Iraq, twelve months after "the surge"? As longtime readers know, the numeraire that this blog uses to measure horror and chaos is the 479 troubles-related deaths in Northern Ireland in 1972 (the worst year of the troubles by a significant margin; the next worst was 1976 with 296 troubles-related deaths). Longtime readers will also recall that we make a genuine request for forgiveness to any Northern Irish readers for doing so; any quantitative discussion of numbers of deaths is always going to be horribly tasteless, but it's also necessary, so sorry. Using a Northern Ireland population of 1.5m, deaths in Iraq have to be scaled by a factor of between 16.5x and 18x to be comparable with Iraq (I'm giving a range here because the last census of Iraq counted 27 million people but there may be as many as 2m refugees living outside Iraq today).

The Iraq Body Count database counts 7974 civilian deaths between 1 January and 22 October for this year. Annualised and scaled (note that I annualised by multiplying by 366/295 which might be too high as the death rate is falling), this would be equivalent to between 1.1 and 1.2 Northern Ireland 1972s.

The month of October (scaled up by 31/22) is the lowest month so far with 462 IBC deaths. That would be equivalent to 0.65-0.7 NI72 units. However, the Northern Ireland figure I'm working off includes British servicemen and both Loyalist and Republican paramilitaries; it's not quite correct to say that the IBC count only includes civilians, but they have quite strict rules for counting Iraqi paramilitaries (basically, only summary extrajudicial executions, although I would guess that more Iraqi paramilitaries are miscategorised as civilians) and don't include coalition deaths at all. Using instead a numeraire of the 250 deaths coded as "Civilian" in the Sutton Index would have Iraq in October at 1.25-1.35 NI72c.

So probably too early to declare victory yet in these terms, although it is worth noting that since June 2008, the level has been below 1.8 NI72, which is worth noting as it is the typical murder rate in Jamaica in a non-election year (last year there were 1574 murders in Jamaica, out of a population of 2.78m). This is still awful, of course, but it's another milestone which shows that the situation is now stabilising to a level of awfulness which is at least comprehensible - in 2006, the IBC database alone recorded 3.2 NI72 units, which is simply unimaginable.

Oh, and of course, there is something of a sting in the end here; from all the surveys carried out in Iraq, from the two Lancet studies to the two UN ones, we have good reason to believe that IBC actually undercounts deaths in Iraq by a factor of between 4x and 10x. Multiply all numbers in this post accordingly, and thank your lucky stars you don't live there.