Saturday, July 24, 2010

Risk and Reward: McChrystal may have been right on civilian casualties in Afghanistan

When GEN McChrystal introduced new, more restrictive rules of engagement in July 2009, reactions were varied. As they were applied in expanding campaigns in southern Afghanistan, though, complaints mounted among troops on the ground:
As levels of violence in Afghanistan climb, there is a palpable and building sense of unease among troops surrounding one of the most confounding questions about how to wage the war: when and how lethal force should be used.
Perennial COINtras like Gian Gentile piled on too, arguing that the emphasis in Afghanistan on avoiding civilian casualties while fighting insurgents is unfounded.

In an ironic twist, on the same day GEN McChrystal ended his career, the National Bureau of Economic Research has released a study that seems to vindicate his approach. Using data from ISAF's Civilian Casualty Tracking Cell, the authors examined the impact of civilian casualties caused by ISAF and insurgents (respectively) between January 2009 and March 2010. Crucially, they examined the impact of those casualties at the district level on both short- and long-term levels of violence, as well as looking for spill-over effects into neighboring districts.

For Afghanistan, the authors found that:

The relationship between civilian casualties and violent incidents in Afghanistan is characterized by three important facts:

(1) There is a positive relationship between civilian casualties and levels of future violence in an area and that relationship is much stronger for ISAF-caused civilian casualties.

(2) Civilian casualties affect the long-run trends in violence, not short-term fluctuations.

(3) The relationship between civilian casualties and violence does not appear to spill over district boundaries.

In particular, they argue that the increased violence following civilian casualties caused by ISAF is the result of increased insurgent recruitment - what they term the revenge mechanism. In contrast, their comparative analysis of Iraq suggests that rather than revenge, Coalition-caused civilian casualties impacted the willingness of the population to share information with counterinsurgents.

The paper's conclusion doesn't beat around the bush:

In sum, the empirical evidence from Afghanistan sheds light on the way in which insurgent groups operate. In particular, it appears that while in high population-density, urban conflicts (such as Iraq) information flows are a critical component to counterinsurgency operations, in more rural insurgencies the most salient factor is the availability of fighters. To the extent that counterinsurgent forces engage in unpopular and aggressive operations that generate specific local grievances, they are likely to facilitate increased recruitment and support for insurgent groups.

In responding to such a situation, military leaders face the task of balancing population protection with restrictions on their own operations. Minimizing counterinsurgents’ harm to civilians appears to minimize the recruiting potential of insurgent forces. The goal of reducing civilian casualties is not necessarily in conflict with the objective of protecting international forces’ lives.

I have yet to fully digest what looks to be an important piece of scholarship and a significant and timely contribution to the ongoing COIN debate. Nonetheless, three questions immediately pop into my head.

  • First, even if inadvertent civilian casualties caused by counterinsurgents do have a measurable impact on levels of insurgent violence, how big is that impact compared to other factors? This may be difficult to measure given the interplay between different drivers, and second- and third-order effects.
  • Second, why (as the study suggests) is the effect of civilian casualties asymmetric in Afghanistan? In other words, why can the insurgents get away with it?
  • Third, how do the different mechanisms and dynamics outlined by the authors regarding the dominant challenges for counterinsurgents relate to the work by other scholars on the collective action problems faced by both insurgents and incumbents?
You can purchase the article from the SSRN (or read what I think is a complete version for free here) and post your own questions in the comments.


  1. MK,

    Thanks for posting this. I’ve been nudging Bill Nagle to get a link up at SWJ for the past three weeks.

    Bottom Line- The five dollars is worth the cost for this article. When Joe Felter speaks, we should listen. He has over twenty years of experience in FID in the Philippines, helped found and coordinate the USMA CTC center, various Irregular Warfare studies, and he is now running the COIN Center in Afghanistan. He will rarely speak to his biography. IMO, he is the quintessential quiet professional that we used to expect from Special Forces officers. Well, I’m not. I’m CAV and proud so I’ll brag on my friends :).

    Our relationship started informally two years ago when I challenged some of his data points in a report that he published. We finally met last year and developed a strong working relationship. He reminds me of the importance of quantitative analysis in conflict study; I remind him that it’s hard to gather surveys at the point of the gun and most of the real information and statistics are hidden from the counter-insurgent during the conflict, and one should be cautious to extrapolate findings just because the research is better than anyone else has done to date. Better is not necessarily accurate. I think our differences make us better, and we have a great personal relationship.

    I would suggest that everyone take the Felter-Few struggle into context with his latest work.

  2. Interesting paper.

    "second, why (as the study suggests) is the effect of civilian casualties asymmetric in Afghanistan? In other words, why can the insurgents get away with it?"

    Does the report actually state this?

    MK, why does a Pashtun Afghan join the ANA and ANP in your opinion? Because they don't like the Taliban? Taliban caused civilian casualties? Or because they want a pay check?

    Why has Pashtun recruitment in the ANA and ANP remained strong during this summer's harvest, when recruiting typically drops?

    Between Nov, 2009, and July, 2009, the ANA grew from about 95 thousand to 134 thousand. The ANP didn't grow as fast because the ANP's 10,300 training base is prioritizing training existing police versus accepting new recruits.

    Be interested in everyone else's perspective as well.

  3. I sense a disturbing narrative forming that is based more in misperception than dispassionate observation. I am not referring just to this blog entry. I am seeing it elsewhere, too.

    First, it seems there is a perception that folks like Col Gentile (or folks like me) who are not under the spell of the COIN spirit reject the notion that avoiding civilian casualties is misguided.

    Second, this study - at least the results of which have been shared above - apparently dispenses with that alleged notion, in a shocking rebuke of the views of we "COINtras."

    To address the first:
    "Perennial COINtras like Gian Gentile piled on too, arguing that the emphasis in Afghanistan on avoiding civilian casualties while fighting insurgents is unfounded."

    That is not a fair or accurate portrayal of what he has been repeatedly and publicly arguing. Quoting from the linked blog entry by Gentile:
    "In short the theory states that by avoiding civilian deaths combined with establishing trusting, emotional relationships the local population can be won over to the government’s side, the insurgents can be separated from the population and either captured, killed, or co-opted, and nation building programs can be put in place."

    It's a little bit more sophisticated than just an opinion on avoiding civilian casualties.

    Second, given that the argument from Gentile et al is a bit more sophisticated than that, this study is not much of a rebuke. Indeed, I don't think the results are a surprise to anyone, whether a member of the pop-COIN religion or those of us who reject such idol worship. ISAF killing civilians causes problems? What a shocker. Largely indigenous insurgents who can easily navigate the society generally get less pushback from killing civilians than the alien occupier? Another shocker.

    But there's another problem here. In spite of all the gibberish I typed above, this study does not vindicate McChrystal. Whether a general policy of avoiding civilian casualties makes sense - that is not really the point. The problem with ROE under McChrystal was not their formulation. It was the implementation. See the now infamous Rolling Stone article:
    The rules handed out here are not what McChrystal intended – they've been distorted as they passed through the chain of command – but knowing that does nothing to lessen the anger of troops on the ground. "Fuck, when I came over here and heard that McChrystal was in charge, I thought we would get our fucking gun on," says Hicks, who has served three tours of combat. "I get COIN. I get all that. McChrystal comes here, explains it, it makes sense. But then he goes away on his bird, and by the time his directives get passed down to us through Big Army, they're all fucked up – either because somebody is trying to cover their ass, or because they just don't understand it themselves. But we're fucking losing this thing."

    Being right is great if you can translate it into success. Being right doesn't do much good if your correct ideas don't make it down to the user level without being distorted.

  4. There's a study at Harvard that apparently is going to refute this study. No, I'm not kidding.