Small Wars Journal

A Framework for thinking about Iraq Strategy

Fri, 01/12/2007 - 1:51pm
The President's new Iraq strategy has prompted much discussion, informed and otherwise. I'm not going to add to it here. Rather, I want to tentatively suggest a framework for thinking about Iraq, which (if you accept its underlying assumptions) might prove helpful in evaluating the new strategy and the enemy's likely response.

I developed this framework about two years ago, while writing the October 2004 version of Countering Global Insurgency, mainly the appendix on Iraq. I have since presented it in various forums, including during the Quadrennial Defense Review in 2004-5, the Eisenhower Series in early 2006, during a series of lectures at the Naval War College and at the State Department's Foreign Service Institute, and during the Irregular Warfare conference in Summer 2006. I also briefed it to the Pentagon's "Plan B" team in November 2006.

So if you were in any of those discussions, read no further since you've heard it all before.

This is a model, not a strategy. That is, it is a systematic oversimplification, designed to clarify an extremely complex, rapidly-changing reality. It does not tell us what to do in Iraq, but is a basis for evaluating options. It is wrong -- all models are -- but applied tentatively, with skepticism, and with constant and rigorous "ground truth" from first-hand observation in theater, I have sometimes found it useful. Here is the model, expressed graphically.

The "Four Problems" concept

In essence, the model suggests that Iraq comprises four strategic problems:

  • an underlying nation-building problem, resulting from the fact that Iraq is a weak and fragile state, and
  • three overlapping security problems that sit "above" that underlying problem, and make it harder to get at it. The three problems are:

    • Terrorism -- that is, the presence of terrorist entities including (but not limited to) AQI who seek to exploit the situation in Iraq to further extremist or trans-national aims
    • Insurgency -- the (primarily Sunni) rebellion against the new post-Saddam order in Iraq, including rebellion against both the coalition presence and the new Iraqi government, and
    • Communal Conflict -- including sectarian conflict between Sunni and Shi'a elements, and ethnic conflict between Kurds, Arabs and other ethnic groups.

These three security problems overlap: incidents may involve elements of more than one dimension -- for example, some terrorism is "pure" AQI activity, while other terrorist acts are insurgent-motivated, and yet others incorporate a sectarian dimension. Most incidents in fact include elements of two dynamics, or all three. You might think of the three problems as a Venn diagram of overlapping circles, each constantly changing in size, with any incident able to be plotted somewhere within the interaction of the three dynamics -- terrorism, insurgency and communal conflict.

They prevent us getting at the underlying problems (crime, weak infrastructure, economic and social alienation, weak governance, an so on) that we need to address in order to deal with the nation-building requirement. The inability to get at this underlying problem perpetuates and exacerbates the security problems.

The three security problems are also mutually reinforcing -- each makes the others worse. Terrorism provokes communal conflict, which in turn makes the insurgency more intractable, which in turn gives rise to terrorism, and so on.

The solution sets to each problem also tend to be countervailing -- the solution to one tends to make the others worse. For example, defeating the insurgency requires building indigenous security forces. But in a society with weak national institutions, divided along sectarian lines, this can tend to make the communal conflict worse. Resolving the communal conflict requires outreach out to all community groups including those (such as some Sunni groups in Anbar) who support the terrorists. But this can create safe havens for terrorists. Countering terrorist cells implies disrupting these safe havens, but that can make the insurgency worse -- and so on, in an endless Iraq do-loop.

Another conundrum is that our presence in Iraq is both essential for a solution, and a source of irritation which tends to exacerbate the situation.

The Regional Dynamic

This whole "four-problem set" sits within a region that straddles a Sunni Arab world (Arabia and the Sunni parts of the Levant) and a Persian Shi'a world (primarily Iran, but also Iranian proxies and allies) with a long history of internal and intra-regional conflict and tension. As General Zinni remarked recently, in toppling Saddam we have created the first Shi'a Arab state in modern history, with profound implications for the long-standing regional dynamic, and the global relationship between Sunni and Shi'a Islam, that we are only just starting to appreciate.

This implies two things. First, conflict within Iraq that threatens to spill into the broader region or drag Sunni states into increased confrontation with Iran, is by definition more strategically ominous than forms of conflict that remain within Iraq. By this measure, the model was already predicting in 2004 that communal conflict would turn out to have broader strategic significance than the insurgency itself (though this tends to change over time).

Second, it is critical that we conduct the campaign in Iraq within a broader regional campaign (diplomatic, economic and informational), rather than conducting it as if Iraq was an island somewhere in the Indian Ocean rather than a central puzzle-piece in a complex regional chessboard. We have to approach the region as a region. This seems like a statement of the obvious, but not everyone seems to have seen the issues this way so far.

Iraq, then, is not a "pure" Insurgency problem

In this sense, Secretary Rumsfeld was perfectly correct when he denied that Iraq was an insurgency and rejected the comparison with Vietnam. This is quite true -- Iraq is not just an insurgency, it is an insurgency plus a terrorist campaign plus a sectarian civil war, sitting on top of a fragile state within a divided region.

And Vietnam is indeed not an apt comparison. The insurgency part of the problem resembles Vietnam to some extent, but insurgency is only one small part of a much bigger problem in Iraq. If we were to draw historical analogies, we might say that the problem in Iraq is like trying to defeat the Viet Cong (insurgency) while simultaneously rebuilding Germany (nation-building), keeping peace in the Balkans (communal conflict) and defeating the IRA (terrorism). And, oh by the way, these all have to be done at the same time, in the same place, and changes in one part of the problem significantly affect the others.

Thus, Iraq is a fiendishly difficult, complex and constantly-changing problem. Look no further for the reason why we have found it so difficult -- put simply, because it is. It's an incredibly complex, tough problem, which requires constant adaptation and agility of response.

(Without getting all theoretical on you, this is what we might call a "wicked" problem, according to the very specific meaning that planners and systems theorists give to that term, in describing a class of problems that has no single solution, no "stopping rule" that tells you when the problem is solved, and where the very act of attempting to solve the problem changes the nature of the problem to be solved. Incidentally, there is a solid and very useful body of research into how to deal with this type of problem, which I have found very helpful in thinking about approaches to Iraq, as have others).

From a practitioner's standpoint, this means that improvements in counterinsurgency technique and capability, while important in addressing the insurgency part of the problem, are not enough to deal with the broader strategic problem in Iraq. Instead, we need solutions that deal with all four problems simultaneously in an integrated fashion, and try to control and impose order on an overall complex environment.

The term I and some of my Australian colleagues developed to describe this form of operation, during work in 2003, was "Control Operations" -- operations that are neither enemy-centric (as in traditional counterterrorism) nor population-centric (as in traditional counterinsurgency) but rather environment-centric, seeking to control and reduce the chaos and violence within a highly complex, multi-sided overall environment. (More about this idea, which I and others have been working on for several years now, in subsequent posts).

So much for theory. In practice, for commanders on the ground, does this help? Not much, I would suggest. The basic techniques of peace operations, counterinsurgency (if executed properly), information operations, counterterrorism, and institution-building are all well-known and at the tactical level our people are well skilled in applying most of them.

Hence the requirement for things like the "Twenty-Eight Articles" -- they don't tell you anything new, nor help address the overall strategic problem (which is well above tactical commanders' pay-grade, thank God), but they help deal with the tactical issues that are already well-understood, and provide memory joggers for guys in the field who have to deal with one discrete part of a larger problem.

Where the model makes a difference is at the strategic level, in prioritizing actions between the various problems, deciding whether and where to expend resources, and -- most important -- developing metrics and "reading the environment" to understand how it changes over time. In a sense, this a prototypical "operational design" for the Iraq theater.

In terms of evaluating options, it allows you to develop checklists or metrics to understand how a plan might function. Does it help us get at the underlying problem? Does it exacerbate the other security problems? Does it address more than one security problem simultaneously? Does it help prevent spill-over of conflict to the regional level? Does it cement state authority? -- and so on. The questions are different each time, but the model helps frame them and work out which factors matter most in a given situation.

Hardly rocket science, of course, just an effort to simplify complex reality and identify macro-trends, which then allow you to work out which bits of detail matter, so that you can drill back down into the detail to think about issues.

Bottom line

I have found this useful as a way of thinking about Iraq and evaluating strategic options. In considering strategic initiatives like those just announced, you need some kind of framework, and this is mine.

If it works for you, use it. If you think of refinements, suggest them (a very senior USN officer whose views I greatly respect did just that in a recent meeting). If it doesn't help you, discard it and think up your own. Either way, you need some kind of model for thinking about Iraq, lest the sheer complexity of events and the difficulty of knowing just exactly what is actually happening overwhelm you.

For what it's worth.....


grl (not verified)

Mon, 02/05/2007 - 7:32pm

Respectfully, I am having a very hard time following your analysis because of how you use labels.

I am not a member of the military, and I am reading this site for the very first time. In no way, shape or form am I an expert on warfare or military strategy.

However, it seems to me that you are using labels usually utilized to describe warfighting methodologies to describe ideological movements which can use any of the warfighting methodologies mentioned to achieve their politcal objectives. This leads to a comparison of fundamentally different things (apples vs. oranges) and a blurring of the definitions of what those things are (apple flavored oranges).

It might achieve greater clarity to separate out the two different classifications so that, instead of three concentric circles, you have a sort of matrix, with war fighting methodilogies lined up on one axis and the relevant political actors lined up on another. (This would also leave room for a third axis, as yet undefined.) You could then do a better job of sketching out how each actor is likely to use each method of warfare.

Just a thought from a non-expert who is trying to get a handle on what you are trying to say.

jheier (not verified)

Wed, 01/24/2007 - 12:19pm

I like the model you've constructed. Do you have a complete list of metrics you'd use for an analysis?


Hi Dave,

Thanks for the responses, and no return fire :).

On the fractal decomposition questions, I was wondering about it for a couple of reasons. I had a feeling that is was, but that was based on reading rather than in the field experience.

On the discourse level, and I'd almost prefer to use memetics as a model for it, I hav a feeling that we could actually use a form of discourse analysis, let's call it "meme tracking", to watch for trends in shifts between the various drivers - basically changes in memetic acceleration or "rate of adoption" and memetic saturation (I wonder if some of the LSI software couldn't be adapted to that). This might give some type of a clue that would allow us to tailor IO ops.

"BUT -- and this is why I am wary of theory -- the hardest thing in counterinsurgency warfare is to know what is actually happening. I don't believe our knowledge on the ground is good enough to justify anything more than a very broad, generic macro-trend set of observations. The more detail you seek, the less clarity you get, and there is a Heisenbourg uncertainty dynamic at work as well."

Granted, and it's always a classic problem, especially with top-down functional models <wry grin>. I think you are also quite right in building this type of model given your audience as well. However, let me put in my own "BUT" as well :)

There is an assumption that runs around in all Western social theory theory that I've read (and I teach this stuff <wry grin>), that seems to state that the theoretical model must be held within the mind of a single individual. This is why the details block the picture - the old "can't see the forest for the trees" metaphor.

But is that necessary for current operations? Honestly, I don't think it is and I think that the indicator for that is that it is not a model used by the radical Islamist movement. We've been having a discussion in the Tactical Bloging thread that touches on this issue, and may provide a partial solution to the problem of too much detail.

"Also, you risk imposing etic categories on local experience."

And that, my friend, is a classic Foucaudian cop-out <GRIN>. Yes, it certainly does happen, no question about it, and it is inevitable. But, as Ginzberg noted, that is an inevitable consequence of and inherent limitation with being human. Since etic categories will always be imposed on any perceptual environment, then the moral imperative, for a social scientist, is to attempt to construct the most accurate etic categories for the specific situation (that was also Merton's rationale for mid-range theory). 'Sides that, isn't this exactly what you have done here <GRIN>?

[in a more serious vein]I'm not saying that you were wrong to do this given your audience - I would have done exactly the same thing and that's also what I teach my students to do. Still and all, I think that other, more situation specific models can be linked into it, maybe using your 28 articles combined with something like tactical blogging, to produce a scalable, robust model that works in many specific situations.

Dave Kilcullen (not verified)

Mon, 01/15/2007 - 4:59pm


I apologize for initially being less than interactive about my first post -- I'm actually traveling overseas and have only periodic access to e-mail. Just a few comments, then, so as not to break the flow of discussion too much:

- on Steve's point about Central America, my impression is that El Salvador (especially the Operation Foundation approach) has some distinct relevance here, although purists would probably dispute the applicability of Foreign Internal Defense models to COIN per se. TX Hammes or others with field experience there might wish to comment....

- On Tom's Lebanon point, couldn't agree more. The similarities with Iraq are extremely interesting (as are the differences). Happy to talk more about that off-line.

- On the points raised in Larry Dunbar's discussion with Prairiepundit and others, I agree with a lot of what has been said. I would just add that in developing the model I considered the growth of local militias (that compete with the government for a monopoly on the legitimate use of force) as a sympton of state fragility rather than as a discrete security problem. They can indeed be a problem, or (provided the state achieves progressive control over them) they can eventually be a force for stabilization, as in post-WW2 Indonesia. So I felt that what we were really looking at here was primarily an outgrowth of state fragility, and explainable that way.

Another major security problem that I chose to include in the "fragile state" category was criminally-motivated violence. Several people have suggested this as a fourth security problem, but I prefer to treat it as a symptom of underlying state weakness.

This is because I designed the five elements in the model (state fragility, international terrorism, insurgency, sectarian conflict and regional instability) as "drivers" rather than physical manifestations. This is a bit theory-heavy, but I was trying to find the fewest number of macro-trends which explained (by themselves or in combination) the largest number of physical incidents on the ground.

-- Now, to Marc's points, I am still a bit unsure on whether the model is fractally decomposable. Clearly, there are fractal-like characteristics and in some areas (Baghdad especially) you can decompose the model through several orders of magnitude and still identify the same five drivers and/or emergent characteristics explainable by their interaction. But then, in other parts (Anbar) only two or three drivers are readily identifiable and a "pure" insurgency model seems to fit better.

On one level, the level of discourse (or memetics) I believe the model is decomposable almost infinitely -- the same issues and the interaction between them are present in virtually all discourse about the war, from the local level to the global.

BUT -- and this is why I am wary of theory -- the hardest thing in counterinsurgency warfare is to know what is actually happening. I don't believe our knowledge on the ground is good enough to justify anything more than a very broad, generic macro-trend set of observations. The more detail you seek, the less clarity you get, and there is a Heisenbourg uncertainty dynamic at work as well. Also, you risk imposing etic categories on local experience.

One final point -- if you look at the very original version of Countering Global Insurgency, in the Iraq appendix you can see that I was playing with influence diagrams and a cyber-feedback theory to explain events. I also developed a biological systems model to try and explain some of the events we see. In the end, I found this type of approach less useful in an Ops Research sense than the more macro-trend model I went with in the end, because although a more elaborate approach is more intellectually rigorous (or seems so) actually you build in a vastly greater set of assumptions and biases into the model, and also create the need to go out and acquire data to support it -- and the act of collection is both dangerous and problematic.... so I lowered my sights a bit.

Also, to be frank, I needed a model simple enough to explain to policy-makers and cybernetics did not cut it....

I'll stop there and wait for return fire before making a couple of other observations....

Dave K

Nice model, Dave. I know you didn't want to get overly theoretical, but I do have a couple of theoretical questions:

1. Do you see the circles as being fractally decomposable and, if so, do you think that could be translated into an ongoing "revisioned" model?

2. How or, rather, where are you accounting for external influences and network ties?

I know you listed AQ in there, which I would count as an external network tie, but I'm afraid that the map might become the territory at the operational level. One of the things that is becoming pretty clear about the overall Islamist movement is that the Muslim Brotherhood is providing clear ideological support and some form of over-arching guidance (at least at the symbolic level).

In answer to Larry, the Shia militia are terrorizing the Sunnis in Baghdad, but if you dig into the story you find their focus is primarily aimed as retribution of the Sunni attacks on Shia non combatants in Baghdad. Many times the bodies are placed on the site of a Sunni car bomb explosion. It is a perverse form of hanging the bodies of coyotes on a fence.

Today, Sadr is keeping his people in a low profile while the Iraqi portion of the surge is underway. I think this supports my theory that he has been operating with he silent consent of the Iraqi government.

As to the question of whether there can be a military solution, of course their can. It is just a matter of whether there is a will to impose one. And, yes, a political solution would be much cheaper in terms of lives and expense. That is why it is the favored path.

As to Prairiepundit's comment:

"While Sadr's thugs may not have an explicit connection to the Maliki government, they are doing the work that the government is incapable of in terrorizing the Sunnis who they believe are backing the terrorist."

Just exactly which Sunnis are they terrorizing? Are they terrorizing the Sunnis in Anbar, where it is reported that the insurgency is, or are they terrorizing the Sunnis trying to live in the mixed neighborhoods of Baghdad. If they are terrorizing the Sunnis of Baghdad, are the Sunnis of Baghdad the same ones supporting the insurgency in Anbar? It is my understanding that the militias are simply ethnic cleansing the mixed neighborhoods of Baghdad. This doesn't seem to me to be where the insurgency resides. It seems to me this is why there is an insurgency.

"While their acts undermine the rule of law, they are not a direct challenge to the government at this point."

I suppose this is true if you believe that there is a military answer to Iraq. But I have not heard anybody say this, and I have listened closely. If there is no military solution to Iraq, then the government of Iraq needs to proportion the oil money to all parties, support minority rights, and form one country called Iraq. It seems to me there is no reason the militias need to even think this way. It is rather like when Arafat was alive. There was no need for him to compromise, because he was going to stay in power either in peace or war. As long as the militia stay as a vertical force inside the Iraqi government there is no need for them to compromise.

In the Middle East there has been a tendency for governments an quasi governments like the Palestinian Authority to spin off groups to do the dirty work that a government is not permitted to do. The al Aqusa Martyr's Brigade was Arafat't terrorist arm. They certainly were not a threat to his rule. While Sadr's thugs may not have an explicit connection to the Maliki government, they are doing the work that the government is incapable of in terrorizing the Sunnis who they believe are backing the terrorist. While their acts undermine the rule of law, they are not a direct challenge to the government at this point. In fact they avoid attacks on government and US forces. I think this, rather than fear of Sadr's political power, is what has restrained the Iraqi government's pursuit to date.

When you assume this rationale it makes sense for the Iraqi government to make its initial focus on the Sunni areas where those who are perpetrating the attacks on non combatants are located, because taking them out will relieve the motivation for the Mahdi army attack.

The problems we have had to date in Baghdad has been our inadequate force to space ratio. This has been primarily the result of poor performance of Iraqi officers who have been left in charge of holding cleared areas. The new strategy adapts the combined action platoon concept to provide some US leadership for the troops. The concept is already working in Anbar and there is no reason why it should not work in Baghdad.

Another interesting aspect of the latest take and hold plan in Baghdad is the gated community concept which has the benefit of giving the troops holding the area a clearly defined space for which they are responsible.

The actions of the militias are counterproductive in the media PR battle because they tend to play into the chaos theme that is a part of the Sunni PR campaign. But, then there has been little effective response to this PR campaign though out the war. The enemy has their script and the media tend to follow it.

<i>Resolving the communal conflict requires outreach out to all community groups including those (such as some Sunni groups in Anbar) who support the terrorists. But this can create safe havens for terrorists.</i>

Perhaps, but resolving communal conflict can also come from victory and defeat. A 15% minority (Sunni Arabs) is able to fight because it has the resources (primarily social and cultural) to do so. <a href="… ethnic militias have been defeated</a> in other parts of the world without this "outreach."

There seems to be another "overlapping security problem" between the government of Iraq and the Shia militia. This Shia militia is also divided up between the Shia more in tune to Persia and that which is more in tune to Arab.

The thing is, this doesn't really seem to me to be an insurgency, as it is well represented inside the Iraqi government. How do you represent these two circles, the Iraqi government and Shia militia (in general)? Are these, then, two overlapping circles and where they overlap you throw in the other three circles? In that way, would you then have to tackle the two outer circles before moving on the inner circles.

Tom Odom (not verified)

Fri, 01/12/2007 - 4:51pm


Thanks for sharing. We used a similar model in the 1990-1991 time frame to explain Iraq, its complexity, Saddam's methods of rule, and effects on the region.

I agree that Vietnam is not the parallel of choice; I look at Lebanon as a more valid case study, one that is not yet complete.

Best regards.

Steve (not verified)

Fri, 01/12/2007 - 3:46pm

Great model to use for Iraq, which is shaping up to be one of the most involved and convoluted Small Wars/CW problems we've seen in the past 50 years. Though I've not dug through the entire corpus of literature, is it possible that there are some parallels to be found in Central America (El Salvador, for one, and possibly Panama) that we may be missing? Columbia might also be a possibility in terms of the level of complexity we're seeing now.