The other side of the COIN

  • Fall 2018
The other side of the COIN
FALL 2018 : précis Student Feature
Matthew Cancian in Afghanistan
Matthew Cancian

Matthew Cancian is a PhD student in security studies and international relations with research interests in military effectiveness and civil wars.

November 28, 2018

The US military’s “can-do” attitude is one of its greatest strengths, but it is also a weakness. Take for example a US Army Major speaking at a think tank conference in 2015. Describing the process of redeploying from Afghanistan to fight Ebola in Liberia, he said without irony, “All we had to do was change the threat vector and we could accomplish the new mission.” The Major’s confidence is certainly an asset among people who might one day be charging into enemy fire, but over the course of the Global War on Terror, such confidence has begun to rub off on scholars of counterinsurgency. It is less useful in this second line of work. These scholars have painted a skewed picture of whether or not counterinsurgency (COIN) is feasible and winnable by relying on shaky theory. They ignore both recent political science arguments about insurgent violence and Carl von Clausewitz’s theory of war, preferring instead to use economic theory that has led to faulty conclusions. Scholars of COIN should be commended for trying to help improve our national security, but the US military should not invest too heavily in their bullish predictions about the feasibility of COIN. Instead, military planners should forswear COIN as a tool of national security.

What is COIN?

Counterinsurgency, or COIN, in its contemporary American incarnation, grew out of the inability of existing US doctrine to contain the insurgency in Iraq. At the start of the Iraq war in 2003, a belief in the revolution in military affairs bolstered by the initial success of the Afghanistan campaign in 2001 led Donald Rumsfeld, secretary of defense, to push for a lighter footprint despite requests from General Tommy Franks, the CENTCOM commander (Gordon and Trainor, 2006). Prior to the invasion, Rumsfeld also refuted Army Chief of Staff General Shinseki’s estimate that over three hundred thousand troops would be needed for the occupation. By 2006, however, high levels of insurgent violence and mounting US casualties prompted civilian leaders to push for a change in strategy. Civilians sought out a maverick in the officer corps (Posen, 1984) and found General Petraeus, who led the re-development of the Army’s counterinsurgency doctrine at Fort Leavenworth and oversaw its first implementation during the Iraq Surge in 2006.

The Army Field Manual (FM) 3-24, product of a commission led by General Petraeus, advocated a ‘population-centric’ approach (as opposed to an enemy-centric one) with the objective of “outgoverning” the insurgents (Department of the Army, 2006). By outgoverning the insurgents, counterinsurgents eventually win the support of the people, who then provide the counterinsurgents with information thereby denying the insurgents anonymity and allowing them to be destroyed. Modern American COIN is not a new doctrine, but rather is based on the experiences of the French and British fighting against communist and nationalist insurgents during the Cold War. David Galula, a French officer who fought in Indochina and Algeria, wrote that “the aim of the war is to gain the support of the population rather than control of territory” (Galula, 1964). British officer Sir Robert Thompson (1966) emphasized the political nature of COIN, arguing that the goal was “to establish and maintain a free, independent and united country which is politically and economically stable and viable.” 

The “can-do” scholars

Taking COIN theory at face value, a decade of microlevel empirical studies using data generated during the American military presence in Iraq and Afghanistan have generally argued that the population-centric model is feasible. Berman et. al (2011) argue, using a formal model and micro-level observational data on Commander’s Emergency Response Program (CERP) fund spending and SIGACTS (records of actions observed by coalition troops), that the provision of services by counterinsurgents does actually lead to increased information provision by locals. In turn, they find this leads to better targeting of insurgents, which leads to decreased attacks on coalition personnel. Beath et al. (2012) use an experiment varying the timing of implementation for a World Bank program in Afghanistan, that improving local governance leads to improved attitudes towards the government. Biddle et al. (2012) find support for the effectiveness of the Surge in decreasing violence in Iraq; Condra and Shapiro (2012) support the emphasis that population-centric COIN places on avoiding civilian casualties, finding that SIGACTs increase following incidents of collateral damage. Another study by Shapiro’s Empirical Studies of Conflict (ESOC) program (Shaver and Shapiro, 2016), replicates the same finding.

On paper, these findings seem to suggest that COIN is feasible because the tactics that it advocates (reducing the use of firepower, using development projects, and increasing contact with civilians) are all correlated with a decrease in attacks on the counterinsurgents. These findings are misleading when it comes to the practicality of COIN.  There are well-known measurement issues with these studies, (for example, attacks on counterinsurgents are not a good proxy for success of a strategy), but there more important and less widely known problems with their models: by accepting the undertheorized FM 3-24 as dogma, these empirical studies of COIN generate a false evaluation of COIN’s feasibility, and this false evaluation has serious policy implications.

It's just a theory

One reason why the premises of population-centric COIN are accepted dogmatically by empirically-minded political scientists is because they interface easily with simplifying rational actor assumptions that are conducive to quantitative and formal modeling methods. These assumptions grow out of the behavioralist setting of political science when the Cold War era COIN authors were writing, combined with the continued acceptance of these assumptions by economists. In the economic sphere of firm behavior, these simplifying assumptions may be valid and illuminating; in the political sphere, however, they can be grossly misleading.Treating insurgents and civilians as instances of Homo Economicus leads to four analytical errors.

First, people have strong ethnic identities that shape their preferences, but COIN theories generally start from the assumption that all civilians are interchangeable. Greenhill and Staniland (2007) and Biddle (2006) both warn modern counterinsurgents against making false analogies between Iraq and Vietnam because those two wars were animated by different ethnic dynamics. On the one hand, Kalyvas’ (2003) emphasis on local conflicts might mean that this was less of a problem in 2003; in the initial stages of civil wars, he argues, much of the fighting is driven by local rivalries that are not related to ethnicity. As the conflict goes on, however, his theory of endogenous conflict argues that the “master cleavage” (in the case of Iraq, ethnicity) grows increasingly important, meaning that by 2006 ethnicity would be a crucial factor, even in his view. 

Second, elites are not epiphenomenal decision aggregators who merely reflect the average preferences of individuals, as they are portrayed in economics, but are in themselves important agents whose role cannot be ignored. This point is illustrated by Christia’s (2012) theory which stipulates that alliance behaviors are based on the interaction of preferences between local elites. Local elites are ignored in economic theories because they cannot be quantitatively modeled; this doesn’t, however, mean that they are epiphenomenal. When I was deployed as a Marine in Helmand province in Afghanistan in 2011, we held a town hall meeting, a shura, every other week per directions from our higher headquarters. Most meetings were attended by around a dozen locals, including a man and his son who had been injured by an airstrike (and whose brother had been detained after a reconnaissance plane saw him serving Chai to the Taliban). At one shura, however, over a hundred people streamed into the base and berated us. Afterwards, we heard that a local landlord had instructed all of his tenants to go and complain on his behalf, possibly because we had driven a convoy over one of his fields. Perhaps we could have won the support of the regular attendee father avoiding collateral damage, but we could not have won over the hundred complaining locals without winning over the local elite, which we never succeeded in doing.

This leads to a third theoretical flaw that has led scholars to argue that COIN is feasible: the idea that economic incentives dominate human decision making in war. Again, this makes for easy modeling, and made sense to early COIN theorists who were fighting communist insurgents; but even communist revolutionary wars were just as often about nationalism and anti-colonialism as they were about redistribution. Authors like Petersen (2002) have written extensively on how perceptions of ethnic hierarchy are important drivers of collective emotions: the loss of status by the Sunnis led them to resent the Shiites in Iraq, leading to conflict that was almost certainly not profit-maximizing. This “technocratic conceit” (Jackson, 2014) induces an upward bias in estimates of COIN’s feasibility.

The final and most serious theoretical flaw in empirical studies of COIN is the tendency to use structural models to describe conflict processes instead of a Clausewitzean, interactional theory. Clausewitz calls war a “Zweikampf” or duel, evoking the image of two wrestlers struggling against one another (Clausewitz, 1976). By this he means that every action in war incites a reaction. Using formal models to determine equilibria does not sufficiently take this interaction into account. SIGACTs might rise in an area because the counterinsurgent is losing and the insurgents have increased freedom of movement, as much existing literature would maintain, but the same empirical signal might also be evidence of the counterinsurgent moving into a hitherto abandoned area and mounting a new challenge to insurgent control. In one story, rising SIGACTs signal failure, in the other, they signal forward progress. The importance of dynamic models is clear in the history of the Uzbin valley in Kabul province, Afghanistan. The Uzbin valley was extremely peaceful until the French rotated in and took over for the Italians in 2008; a month later, a massive ambush killed 12 French soldiers as they patrolled into a new area. This is not to say the Italians had been “winning” and that the French had begun to “lose.” Rather, Italians had been paying the Taliban to keep quiet while they huddled in their FOB and had not told the French: the level of SIGACTs had no correlation to COIN success. To bring it back to an example in the political science literature, Sexton (2016) uses the presence of American Forward Operating Bases (FOBs) to proxy for secure coalition territorial control. This is a mistake. It is true that the United States does not have FOBs in areas of Taliban control; they also don’t exist in the quietest areas of the country where they would serve no purpose. In ignoring the interactional nature of war and modeling it as a product of structural factors, empirical proponents of COIN feasibility mistake war’s fundamental nature and thus make false inferences about COIN’s feasibility.

The future of COIN: Best left in the past

On last year’s SSP trip, our program had the opportunity to visit the Army’s National Training Center in Fort Irwin, where the Army prepares brigades for future conflicts. There was some talk of “full-spectrum warfare”, the concept from the 1990s that an Army optimized for conventional conflict could also do COIN; the fact that these conflicts don’t exist on a spectrum but are two entirely different phenomena was clearly revealed by our experiences during the Global War on Terror. Thankfully, at Fort Irwin, the focus was on defeating Russian mechanized divisions. While there was a village where some notional COIN work was practiced, the focus was decisively on the conventional battles that preceded and followed the occupation of the village. It seems that the Army has learned the right lesson from Iraq and Afghanistan: COIN is probably beyond the capabilities of our national resolve and best left to the history books. Whether politicians and scholars will draw the same lesson, however, remains to be seen.



Beath, Andrew, Fotini Christia, and Ruben Enikolopov. "Winning Hearts and Minds through Development Aid: Evidence from a Field Experiment in Afghanistan." Centre for Economic and Financial Research at New Economic School Working Paper 166 (2012).

Berman, Eli, Jacob N. Shapiro, and Joseph H. Felter. "Can hearts and minds be bought? The economics of counterinsurgency in Iraq." Journal of Political Economy 119, no. 4 (2011): 766-819.

Biddle, Stephen. “Seeing Baghdad, Thinking Saigon.” Foreign Affairs, (2006): 2-14.

Biddle, Stephen, Jeffrey A. Friedman, and Jacob N. Shapiro. "Testing the surge: Why did violence decline in Iraq in 2007?" International Security 37, no. 1 (2012): 7-40.

Christia, Fotini. Alliance Formation in Civil Wars. Cambridge: Cambridge University Press, 2012.

Clausewitz, Carl von. On war. Trans. Michael Howard and Peter Paret. Princeton: Princeton University Press, 1976.

Condra, Luke N., and Jacob N. Shapiro. "Who takes the blame? The strategic effects of collateral damage." American Journal of Political Science 56, no. 1 (2012): 167-187.

Department of the Army. "FM 3-24: Counterinsurgency." (2006).

Galula, David. Counterinsurgency Warfare: Theory and Practice. Westport, CT: Greenwood Publishing Group, 2006.

Gordon, Michael R., and Bernard E. Trainor. Cobra II: The Inside Story of the Invasion and Occupation of Iraq. New York: Vintage, 2006.

Greenhill, Kelly M., and Paul Staniland. "Ten ways to lose at counterinsurgency." Civil Wars 9, no. 4 (2007): 402-419.

Jackson, Colin. "Government in a box? Counter-insurgency, state building, and the technocratic conceit." In The New Counter-insurgency Era in Critical Perspective, pp. 82-110. London: Palgrave Macmillan, 2014.

Kalyvas, Stathis N. ‘The ontology of “political violence”: action and identity in civil wars.’ Perspectives on Politics 1, no. 3 (2003): 475-494.

Petersen, Roger D. Understanding Ethnic Violence: Fear, Hatred, and Resentment in Twentieth-Century Eastern Europe. Cambridge: Cambridge University Press, 2002.

Sexton, Renard. "Aid as a tool against insurgency: Evidence from contested and controlled territory in Afghanistan." American Political Science Review 110, no. 4 (2016): 731-749.

Shaver, Andrew, and Jacob N. Shapiro. "The Effect of Civilian Casualties on Wartime Informing: Evidence from the Iraq War," HiCN Working Papers 210 (2016), Households in Conflict Network.

Thompson, Sir Robert. Defeating Communist Insurgency: The Lessons of Malaya and Vietnam. Vol. 10. New York: FA Praeger, 1966.

Posen, Barry. The Sources of Military Doctrine: France, Britain, and Germany Between the World Wars. Ithaca: Cornell University Press, 1986.