Wednesday 06 Oct 2021 | 14:59 | SYDNEY
Wednesday 06 Oct 2021 | 14:59 | SYDNEY

Unpacking the data on counter-insurgency

4 April 2011 14:17

I would like to respond to John Hardy's four objections to my argument. John's first objection is that disregarding the unresolved conflicts skews the trend data. This is wrong.

Regardless of the number of unresolved conflicts, there are still more counter-insurgent wins after 11 years than there are after five years. The 'omissions' do not, and cannot, change this maths.

John's second objection (are pauses in conflict 'wins'') has more merit. However, does it demand that we consider the current International/Afghan — Taliban War as an extension of the Soviet — Afghan War' Clearly not. John points to the Turko — Kurdish dispute as an example, and I don't claim any knowledge of this conflict.

The authors of 'A Thousand Fathers' laid out clear guidelines for what constituted an active insurgency and what conditions determine an insurgent or counter-insurgent win.

Turkey has met the conditions of a win and I assume that the current level of any Kurdish insurgency does not reach the threshold to be considered in the data. The study methodology is clear, well articulated and consistent. John's second objection may have merit, but we are discussing someone else's dataset and analysis, and there is no suggestion of serious methodological error in the study.

The trend remains, albeit at a slightly weaker level, if we adjust the data to take account of John's objection. Like the above, this does not materially change the numbers even if we completely accept the most extreme extension of John's argument, that all three Afghan wars are the same conflict (which I don't believe John is arguing). If we absent this extension, then there is no change to the trend.

John's third objection is that only three of the thirty cases included an external actor as the prime counter-insurgency (COIN) actor. This is noted in the study, however a careful reading shows that the counter-insurgent must have more good practices than bad to succeed. Whilst an external COIN actor is not necessarily good or best practice, it is also not terminal. We should recall that in a number of these conflicts, there was significant external assistance in favour of the insurgent.

This assistance was probably decisive in the cases of Kosovo and Bosnia. We should also recall that the mission in Afghanistan is to hand over security responsibility to a legitimate and stable Afghan Government. On this point, significant arguments can be made against continuing our involvement in Afghanistan. 

The fourth objection about control of the border is pretty much the same as the response to the third. I first noticed the trend on time at the beginning of the year and posted about it on the Small Wars Journal. This observation raised some eyebrows, as it is counter to conventional wisdom. On Thursday I read Gorka and Kilcullen's article 'An Actor-Centric Theory of War'. In this article Gorka and Kilcullen use the correlates of war database to assess recorded conflicts between state and non-state actors from 1816 to 2007.

The authors state that: 'Surprisingly, despite the conventional wisdom, in 80% of conflicts, the government defeated its irregular foe' In my paper on 'Counter-insurgency: Domestic Politics by Other Means' I argued that insurgency always reflects a condition of political weakness. Either ideological weakness, where the insurgent can't convince the population to voluntarily join its cause and therefore must coerce the population into support; or structural weakness, where there is no legitimate space for groups to contest domestic political power. The weakness of the insurgent contrasts to the great strengths of the government.

This point is critical, because in both Iraq and Afghanistan we have sought to create the opportunity for different groups to peacefully and legitimately compete for domestic political power — which is what the conflicts are about. We should therefore assess the position of the insurgent as ideologically weak, a point highlighted by most polling on popular Afghan attitudes to the Taliban.

However, we must be careful not to confuse ideological weakness with political will to fight, and of the two, will is more important. Given the previous concern for statistically valid sample sizes and the ability to confidently generalise, Gorka and Kilcullen note that: 'Additional data show that on average, the successful counterinsurgent will need 12 to 15 years to defeat an insurgency. According to several studies, those insurgencies that defeat government do so in 5 to 9 years.'

I believe this larger data set from eminent scholars substantially validates my initial argument that time is an important factor in counter-insurgency and the longer the conflict continues, the more likely the counter-insurgent is to win.

However, as I have argued all along, time is but one factor and good practice is the strongest determinant of success. I believe that my posts have demonstrated that the historical evidence clearly supports my argument and no counter argument supported by compelling evidence has yet been made.

Photo, of former US General Stanley McChyrstal, by Flickr user isafmedia.