• About
  • Categories
  • Features
  • Resources


Add to Technorati Favorites



« Replacement Geography - Get Off Yer Fat One | Main | Spiking the Meme »

Modeling Afghan Power Structures

You know what a qawm is, right? It’s an Afghan solidarity/patronage network and it’s a pretty important aspect of Afghan society. It’s a commonly enough analyzed concept. You can find a qawm bibliography at this end of Afghanistanica’s discussion of qawm. But what I’ll deal with today is Armando Geller and Scott Moss’s computational model for qawm relations.

Geller, Armando and Scott Moss. 2008. ‘Growing Qawm: An Evidence-Driven Declarative Model of Afghan Power Structures’, Advances in Complex Systems, Vol. 11, No. 2: 321-335. Download PDF here.

Geller and Moss conclude that:

Mutual interaction among different types of actors leads to the emergence of a complex organizational structure which contains a number of centers of power, which themselves consist of a number of political, economic and military stakeholders cooperating in a limited way. That these networks tend to be small-world in scale has an important politico-social implication: because smallworld networks are regarded as being more robust against external perturbances than alternative network topologies, this research suggests that qawm exhibit a high level of robustness toward exogenous influences.

But the authors also blame the prevalence of qawm structures for Afghanistan being a rentier economy, as a small number of qawm power hubs accumulate a grossly disproportionate share of local resources. In addition they cite qawm “loyalties” as a reason for the fragmented nature of Afghan society.

But the authors dedicate most of case study to running computer simulations of the evolution of power structures and the actions of the players within them.

In the Afghan context, the interactions within qawm determine a pattern of actions that could be described as episodic clusters of cohesion building and dissolution. Such qawm level behavior leads to interactions amongst the qawm that cause episodic alliance building or conflict of unpredictable magnitude, duration and outcome.

The network, at the elite level, looks something like this:


Geller and Moss note that forecasting is not a possibility, but they do reach these findings:

We find that neopatrimonial behavior is a highly socialized strategy to gain power. Isolated actors are doomed to fail — even if they would have the personal potential to become powerful. Moreover, neopatrimonial behavior requires constant activity. Actors must be continuously aware of potential chances to affiliate themselves with other actors for their own good. However, the analysis of the dynamics of the endorsement process also reveals that neopatrimonial behavior leads to the emergence of a segregated society. The simulated neopatrimonial behavior also produced a macropattern that resembles the dynamics prevalent in conflict-torn societies, where all-out war is the exception, but constant noise and hard-to-forecast volatility in terms of the timing of an event and its magnitude is the rule. While such a pattern is characteristic of Afghanistan and contemporary conflicts in general, it is also an indicator of the existence of dynamics found in complex systems.

This is the most important point that they made:

The simulation results suggest that the emergence of qawm and, hence, the fragmentation of Afghan society are systemic and lead to a constant drain of resources.

This is the only point where I have a problem. At the lowest level, where qawm affiliations in rural areas may be a necessity for survival and where the relationship is more personal, even kin-based, I’m not convinced that the qawm structures lead to a draining of resources in most cases. Though I should mention that Geller and Moss explicitly note that there study is based on interviews conducted in urban areas and their model focuses on “strongmen,” with whom the process of accumulation and distribution of resources better fits a pattern of resources being drained. So I’m in agreement with Geller and Moss that “Upcoming research must scrutinize competition and conflict in qawm-like power systems more intensively.”

But honestly, to give this article an honest critique, one probably needs a background in sociological computational modeling. I might be misinterpreting some of what Geller and Moss are trying to convey.

Download the article here.

Read Afghanistanica’s discussion of qawm here.

PrintView Printer Friendly Version

EmailEmail Article to Friend

Reader Comments (12)

Christian, interesting that you should cover this. I'm sharing a panel with Geller at the British International Studies Association meeting at the University of Exeter in December. See here.

See also London's Centre for Policy Modelling, Geller's home base.

Jul 5, 2008 at 10:02 | Unregistered CommenterMike

The merits of this paper by Geller and Moss to their discipline can be evaluated by their academic peers. It does not contribute to the effort to infuse those who are operating in the field and affecting the lives of Afghans with the qawm concept and the perspective that it brings. For that, Afghanistica's brief paper is an effective tonic for opening the eyes of those who are willing to deal with reality's complexity and uncertainty.

Jul 6, 2008 at 0:36 | Unregistered CommenterDavid

Dear all,

Thank you for your interest in our paper. Allow me to make some brief remarks.

I agree with Christian's comment that he is "not convinced that the qawm structures lead to a draining of resources in most cases." However, most cases in Afghanistan are those cases that do not matter with regard to overall resources. Material resources in Afghanistan are so unevenly distributed that the small number of cases where the idea of "draining resources" (or rentier economy) applies to have a huge impact on the whole economy. But I am very grateful for the comment, as I do not want to imply a generally negatively connotated notion of qawm.

I do not agree with Christian's comment that someone should be into modelling him/herself to critically comment on the article. The sort of modelling we are promoting is not situated in the abstract -- it wants to be part of a case-study-based and policy oriented research agenda. It wants to be understood as a complementary approach to more orthodox qualitative and quantitative research designs. Hence, I could not disagree more with David's black and white account to research approaches.


Jul 6, 2008 at 10:09 | Unregistered CommenterArmando Geller

Fair enough. I do agree that there is an overall draining of resources due to a minority of stakeholders controlling and very unevenly distributing the majority of resources. And this will, of course, end up affecting most everybody as it harms the whole economy.

Jul 7, 2008 at 11:59 | Registered CommenterChristian Bleuer

Armando, this is a very interesting article, and I find the points you make about the network structure fascinating. I do have one question, however, on where you note, "Actors must be continuously aware of potential chances to affiliate themselves with other actors for their own good." Have you examined the literature on both 'preferential attachment' and/or 'balance theory'?

My suspicion is that part of what you are observing might be a bi-product of these network dynamics, although your networks may be too small to truly test for preferential attachment.

Jul 7, 2008 at 19:58 | Unregistered CommenterDrew Conway

Christian, unfortunately this is the situation as I see it. The question therefore remains: How to work together with those people that are in power and hold most of the country's assets in order to create a situation that is beneficiary not only to a small elite? I briefly touch on this problem in a book chapter that should be published by this fall by Hurst & Co. Antonio Giustozzi thinks about these issues too in his article "War and Peace Economies of Afghanistan‟s Strongmen" (International Peacekeeping 14, 1 (2007), pp. 75–89). Finally I know of other researchers and policy makers who started to reason in these terms. Although ethically and implementationally challenging, I see it currently as the only feasible procedure.

Drew, thank you very much for your comments and suggestions. They are very helpful. To be frank: No, I have not thought about preferential attachment and/or balance theory. But now that you made me aware of it, I will. Thanks!

Jul 8, 2008 at 16:16 | Unregistered CommenterArmando Geller

Being neither a network theorist nor an Afghanistan specialist, I'd be interested to know more about how qawm dynamics potentially shape or enable new security and militant structures. The social constructivist implications are readily apparent, and run counter to what one might normally assume: that ethnic and tribal dynamics trump all, including the potential to build or develop alternative loyalties or morph into new cultural formulations. That, obviously, isn't the case. So what do qawm dynamics suggest for security sector reform, on the one hand, and insurgent and terrorist activity, on the other - where the strength of organizational culture is integral to its ability to function?

Jul 9, 2008 at 11:20 | Unregistered CommenterMike

The qawm dynamics within the security and military structures are definitely worth looking at (even if it is hard to categorize categories that are so fluid, even compared to other categories). A good example was Rashid Dostum's militia. In the early 1980s he recruited on an individual basis and formed a very modern looking military force that was very capable. But he later turned to recruiting qawm leaders, who entered with all of their followers. While great for the recruiting numbers, it was very bad for command and control. And he paid the price for this in '97.

Qawm dynamics can show a variety of patterns. It has been claimed that Uzbeks will follow their Qawm leader while Pashtuns will not do so as easily. The Taliban seemed to realize this and recruited on both an individual and elite basis while eliminating those leaders they knew would continue their "bad" ways. They basically targeted the problematic leaders while offering incentives to their followers to join. US/NATO/ISAF/GOA seemed to have made a mistake early on and recruited and integrated many of these bad commanders at the same time they did their followers. This has been somewhat remedied, maybe.

As for the insurgency and terrorism, recognizing the reasons why some Afghan Pashtuns support or assist the Taliban has much to do with the competitive dynamics at a local level. But as for terrorism, I don't see any qawm dynamics at play.

Jul 11, 2008 at 7:56 | Registered CommenterChristian Bleuer

I should clarify my meaning when I refer to insurgency, terrorism and organizational robustness.

What I had in mind was Bruce Hoffman's modeling of non-state armed groups. In Inside Terrorism (rev. ed., p. 35-38), he distinguishes terrorism from other forms of politically violent non-state groups, as an aid to defining the problem. In this construct, he sets up a typology of guerrilla, insurgent, and terrorist formations, along with a brief discussion of criminal groups.

In this, guerrilla formations are larger in number and territorially concentrated. Insurgents are somewhat less so, embedding elements of their organizations within the larger population. Terrorists are cell-based, immersed in the general population and largely indistinguishable from it.

The overall argument is much more nuanced, and of course I don't do it justice here. Its ideal types aren't universally applicable, which Hoffman himself notes. My point in raising the issue is simply to suggest that various forms of armed militancy have specific and evolving spatial requirements and characteristics.

In light of this, and given the apparently strong locational dimension of qawm, does it or can it relate to armed non-state groups? Setting aside for a moment the obvious differences in motivation, intent, and conduct that distinguish guerrillas, insurgents, terrorists, and criminal organizations, how might something like qawm shape or inform any of them? How might qawm interact with various scales of armed militancy? Is there an element of qawm that might contribute to the resilience of armed militancy? Is there an element of qawm that might facilitate pan-tribal, pan-ethnic, or pan-motivational associations?

Jul 11, 2008 at 18:51 | Unregistered CommenterMike

Hmmm. I guess I shouldn’t have sold that book back to the university book store for $4 back in the day, but that buys you lunch in Indiana.

Short answers: I lean towards qawm assisting in decentralized militancy but being an organizational liability when it comes time for an organization to centralize and better coordinate its activities as it hopes to transform into a government, even at a regional level, or even just move beyond the initial stage of conflict. Qawm politics are highly focused on local issues and are, in my opinion, a liability to organizational strategy and control when it comes to pan-whatever. But when the leadership is vulnerable and under high-pressure (or even out of theatre) the ability of qawm units to operate autonomously and with high motivation is an asset.

But that being said, qawm in its broadest definitional is so broad as to be nearly meaningless and qawm dynamics can manifest themselves in different patterns throughout Afghanistan.

Jul 12, 2008 at 9:01 | Registered CommenterChristian Bleuer

Dear All, I apologise for my utterly late reply.

Mike: One important lesson from a "qawm-based" approach, I think, is that security sector reform -- as any other reform in Afghanistan -- is in need of contextualisation. The notion of qawm highlights Afghans' multiple social identities and any kind of social engineering (please forgive me the politically incorrect expression) needs to take this fact into consideration. In oder words: A viable and sustainable reform of the security sector must respect Afghan socio-political traditions and include them in the re-organisation of the security sector. With regard to insurgent and terrorist activity the idea of qawm tells us that there is no such thing as fighting "the" Taliban. The insurgency (alternatively: resistance) in the South is a heterogeneous phenomena and the groups representing it are amorph and opportunistic.

Christian: I agree with you that qawm dynamics do not play an important rôle in the case of terrorism, especially suicide terrorism, as coercing individuals into terrorism is an important recruiting "argument". And I do also agree with you that local dynamics are of importance in groups' decision whether to collaborate with/join the Taliban or not. But the notion of qawm is particularly essential on the local level, where various identities overlap, such as family, village, valley, etc. I disagree, however, with your idea of leadership. Although commanders, such as Dostum, but also Ismael Khan and Ahmad Shah Massoud, were accepted as leaders, they always had to be aware of centrifugal forces within their organisations (Jamiat-e Islami, Jumbesh-i-Milli). This danger particularly stemmed from ambitious sub-commanders. So the organisational incorporation of different qawm into a larger qawm -- in this case a politico-cum-military endeavor -- is part of Afghanistan's recurrent political reality and all of these leaders could not and still cannot think of political life without it.

Mike: The last 30 years in Afghanistan have been also characterised by an import of socio-military behaviour that was before inexistent in Afghanistan. In the 1980s Afghans have been, for example, trained by Pakistani and the British in modern guerrilla warfare; Arabs, for example, have introduced them to religious extremism and, more recently, as evidence suggests, to suicide terrorism. While the former has certainly merged with the idea of qawm, the latter has not (see above). Thus, I think that qawm can relate to modern ideas brought to Afghanistan. However, in the case of Bruce Hoffman's typology, as is with any other typology, the question should not be if the observation fits the typology, but if the typology can grasp the observation. But back to your original question (and to Christian's statement, 12 July): Yes I think that the idea of qawm contributes to the resilience of armed militancy and some kind of "pan-ism". The qawm's importance for these questions does not lie in the fact of identities' multiple idiographies, but in the fact that it allows its members to think and (successfully) act in a highly dynamic socio-political environment. This is reflected in concomitant multiple loyalties and the immediate capitalisation of opportunities. Hence, answers to your question are to be found in qawm internal mechanisms rather than in their actual manifold structural materialisations.

(P.S. Christian: What book are you talking about?)

Aug 11, 2008 at 10:45 | Unregistered CommenterArmando Geller


The book Chris mentioned was a reply to my reference to Hoffman's Inside Terrorism, from which I drew the spatial typology of guerrilla/insurgent/terrorist organizational requirements.

Aug 12, 2008 at 4:24 | Registered CommenterMike Innes

PostPost a New Comment

Enter your information below to add a new comment.

My response is on my own website »
Author Email (optional):
Author URL (optional):
Some HTML allowed: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <code> <em> <i> <strike> <strong>