It is well established that this summer saw an upwing in violence in Iraq. August saw the largest number of deaths of Iraqi civilians from violence since the beginning of Operation Iraqi Freedom. While there was a decline in both September and October, such things are not news. The August upswing was firmly on people's minds when America went to the polls in November. It is a key reason why Nancy Pelosi is Speaker of the House. I just have the one question. Did it actually happen?
The figures speak loudly. These are the numbers of Iraqi civilians killed by violence for each of the last six months of 2006, excluding ordinary civilian murder:
748, 844, 2589, 1150, 560, 826, 524
As you will see, the third figure, August, shows the huge upswing we have been talking about.
What, then, are these:
2401, 2746, 420, 2195, 3149, 2816, 2390
Those are the figures for ordinary civilian murders in Iraq for the same six months. Look at August. An equally sudden and equally unexplained downswing that precisely mirrors the upswing for deaths attributed to insurgency, terrorism and sectarian strife. For the sake of clarity (and to avoid the Brookings Institute's policy of saying one set of figures includes murders when I consider them all to be murders) I will call one set the political killings and the other the apolitical killings.
Is it credible that an increase in political killings should be almost precisely balanced by a decrease in apolitical killings? Would it not be fairer to look at all killings and try to discern a trend from that? Okay, here goes:
3149, 3590, 3009, 3345, 3709, 3642, 2914
It is a fairly stable set of figures. Not only is there no upswing in August, August is actually slightly down on the months either side and is the second lowest figure of the set.
Okay, then, is it possible that something happened which suddenly, and temporarily, radicalised Iraqi youth? Something which sucked them away from killing their spouses, difficult neighbours and victims of their streetcrimes and persuaded them to engage in political killing instead?
It IS possible. In a fascinating article in the FT (Link for subsribers) Christopher Caldwell points out that there are 67 countries or territories around the world where people aged 15-29 comprise more than 30% of the population. Sixty of them (including Gaza, which his article is mostly about) are engaged in some form of civic unrest or civil war. The demography of Iraq, where the median age is around 19, is not as extreme as Gaza, where it is about 15, but is still severe.
Looked at on that basis it is possible that countries have their own natural level of violence, set by demographic and cultural factors, and political factors merely help define its pattern. Perhaps political events - and the Israel-Hizbollah clash of the summer seems the most likely culprit - really did suck people away from their ordinary apolitical killings into political killings.
But the changes in August of 2006 are too sudden, too large, and too temporary to smell right. It seems far more likely that there was a counting error. Marginal cases which might otherwise have been counted as apolitical were, in August, counted as political. I have no explanation as to why this might be so, except perhaps that media coverage of the violence influenced people engaged in the count.
So what have we established? First, there was no overall surge in the levels of violence. That simply didn't happen. Second, the figures which show a shift from apolitical to political violence are deeply suspect.