Wednesday, May 18, 2011

The legend of COIN as essential CT enabler: the bin Laden exemplum

Spencer wrote a thoughtful post last week on a question I'm surprised hadn't been brought up earlier: does the bin Laden raid tell us anything about the legitimacy of past claims that a large-footprint counterinsurgent force on the ground in Afghanistan is essential to collecting the kind of solid, reliable local intelligence we need to wage effective operations against al-Qaeda? Take it away, Attackerman:
So the bin Laden raid leaves us with an empirical test case for a core counterinsurgency contention.
Proposed by Dave Kilcullen and others, counterinsurgency theorists argue that theirs is the truer path to robust counterterrorism, the pursuit of which is the only reason counterinsurgency is germane in the U.S. context. Counterterrorism requires intelligence on terrorists and their environs. Intelligence requires local knowledge. Local knowledge requires the protection of vulnerable populations from reprisal, plus some incentive -- money, justice, political access, whatever -- to make cooperation with the anti-insurgent faction sensible. If you don't provide those resources and meet people where they are, then your counterterrorism hunt is fruitless. 
U.S. counterterrorism just succeeded, massively, in a place where we don't have a counterinsurgency campaign. But we clearly had sufficient local knowledge (read: informers and ground operatives) -- plus signals intelligence, plus persistant surveillance -- to meet our objective.
This seems like a pretty simple deal, right? Let's reformulate what Spencer has written in the simplest possible terms:
  1. Some COIN advocates sold their operational approach by saying that it involves the collection of intelligence that enables accomplishment of the primary national objective, which is the destruction of al-Qaeda and its associated movements. 

  2. A significant step towards the accomplishment of that objective was achieved in a manner that suggests that the COIN campaign in Afghanistan did not meaningfully contribute. 

  3. As such, we should reconsider whether the massive expense, reduced operational readiness, and other sacrifices being made to enable that COIN campaign are actually essential to accomplishment of the fundamental goal; i.e. is the war in Afghanistan really accomplishing anything vis-a-vis the president's stated objective, which is, lest we forget, to "disrupt, dismantle, and defeat al Qaeda"?
This isn't going to come as a shock to you, but I completely agree with all of that. I agree with the line of thinking. I agree with the need for re-assessment. (To be more accurate, I think the assessment -- whether or not our ways and means can meaningfully contribute to the accomplishment of our ends -- should have been done and then explained and justified to the American people before the escalation too place, because it's not as if these concerns are new (pdf)... but we all know how that went.) I agree that this should serve as a kick in the ass for policymakers to genuinely think about why we're fighting the war in Afghanistan and what benefits we're reaping in exchange for its various costs. Agree with all of that. But the original post actually requires a little bit of semantic parsing, because the COIN pitch made by escalation advocates was multi-layered, sophisticated, and not vulnerable to a single point of failure. So let's take a second to get some details straight.

During the debate over possible courses of action that preceded the president's announcement of an escalation, the various positions were basically boiled down to "CT," "CT-plus," and "COIN." I'm not going to rehash the details of all of that, because it's boring and you almost certainly already know the broad strokes. (If not, read this.) But!: one of many, many insidious effects of that simplification was the way that "counterterrorism" or "CT" came in the popular lexicon to mean something like targeted direct action against terrorist networks, personnel, and support infrastructure -- i.e. killing terrorists -- without what constitutes the other half of that definition in military doctrine: "action taken indirectly to influence and render global and regional environments inhospitable to terrorist networks." Now, it may be the case that the DOD definition is too broad to be meaningful, and you could certainly make that argument. But by this definition and according to the justification of the mission provided by Presidents Bush and Obama, any USG activities undertaken in Afghanistan would constitute "counterterrorism"... including COIN, FID, SFA, development assistance, economic support, etc etc ad infinitum.

Now take another look at Spencer's post in this light. "Counterinsurgency theorists argue that theirs is the truer path to robust counterterrorism, the pursuit of which is the only reason counterinsurgency is germane in the U.S. context." He goes on to explain the presumed linkage between security to governmental capacity to popular loyalty to local knowledge to actionable intelligence, which serves to clarify what he really meant by "the truer path to robust counterterrorism": COIN ops will presumably result in military forces getting more information about the environment and the enemy than they otherwise would have (without sustained on-the-ground presence), meaning you can kill more terrorists this way.

I want to vigorously dispute Spencer's characterization of this questionable (and opportunistic) corollary as "a core counterinsurgency contention." What we see here is a cunning conflation of two superficially plausible semi-truths: 1) persistent, visible, dispersed presence by counterinsurgent security forces will enable the accumulation of local knowledge that is essential to effectively targeting insurgent cadres and thus enabling the broader counterinsurgency effort, and 2) the local knowledge that's accumulated as a result of sustained presence and security will include intelligence that allows the counterinsurgent force to undertake targeted, effective offensive operations against al-Qaeda terrorists (whose numbers in the Afghan AO are infinitesimal if not zero).

It's only fair to further note that when Spencer writes "the pursuit of [counterterrorism] is the only reason counterinsurgency is germane in the U.S. context," he's engaging in a bit of selective memory (or revisionist history). Many advocates for escalation and a broad-based COIN effort in Afghanistan were proponents of a widely-held view in both liberal internationalist and neoconservative circles, one that has only recently -- with a downturn in popular support and enthusiasm for overseas adventurism -- started to recede from the public forum: faith in what we can call the "state strength" paradigm. They believe(d?) that weak states posed the greatest threat to global and national security, that poverty, disenfranchisement and lack of economic opportunity not only drove local conflicts, but threatened Americans' way of life. Here's how it's pitched (pdf):
We know where extremists thrive. In conflict zones that are incubators of resentment and anarchy. In weak states that cannot control their borders or territory, or meet the basic needs of their people. From Africa to central Asia to the Pacific Rim – nearly 60 countries stand on the brink of conflict or collapse. The extremists encourage the exploitation of these hopeless places on their hate-filled websites.
In this telling, poor economic conditions and injustice stoke disaffection and rage while weak governments are insufficiently capable of taking steps to either ameliorate these conditions or control their disaffected populations, and the whole thing ends up with pissed-off foreigners flying airplanes into office buildings. In short, they believed that failed states created terrorists, and that the best way to fix these failed states would be to build governmental capacity through a wide-ranging nation assistance and state-building effort while taking direct military action to defeat the most immediately dangerous security threats... what's been called, in the imprecise shorthand of our political narrative, "COIN."

In a defense and foreign-policy establishment where this worldview predominates, almost to the exclusion of all others, is it fair to say that "the pursuit of [robust counterterrorism] is the only reason counterinsurgency is germane in the U.S. context"? I would argue that it isn't the only reason, and that Kilcullen et al's focus on the CT-enabler angle was merely a matter of opportunistic calculation to appeal to those few "realists" who may not have fully bought into the contemporary conventional wisdom. The real challenge doesn't come from such obviously inadequate policy justifications as the narrowly-circumscribed one that Spencer seeks to highlight, but rather from the undifferentiated monolith that is the state-strength paradigm. It is close to a fact of political science that the risk of civil violence is closely correlated with low per-capita GDP/quality of life (as Jay Ulfelder points out today in a fortuitously-timed post). But it is far more difficult to establish causation, and causation is what we need to understand if we want to bring policy instruments to bear on a security problem. Consider just a few of the multivariate assumptions about the causes of conflict and violence that are embedded in the state-strength paradigm.
Violence is caused by: poverty; unequal access to or distribution of resources; unemployment; the inability of the justice system to fairly adjudicate disputes; the inability of the police to enforce the writ of the state; the inability of security forces to maintain the monopoly of violence; the inability of security forces to defend borders/sovereignty/protect against outside influence and/or infiltration of adversarial elements; the inability of security forces to undertake and/or sustain effective offensive operations against identified targets such as terrorists or insurgent groups; etc.
Some of these may be true, some specious, and some outright absurd. But the lumping of these many purported causes, drivers, or enablers of conflict under one broad rubric -- "state weakness/failure" -- has allowed for that paradigm's furtherance through unified (or at least complementary) advocacy by disparate power bases and interest groups: DoD can talk about the requirement to build capacity and capabilities in partner security forces in order to defend borders and eliminate safe havens; while State can emphasize the need to build whole-of-government capacity with a specific focus on the institutions of justice and so on; and USAID and the development community can pitch the imperative of improving the delivery of basic assistance to address human suffering, economic inequality, and other purportedly fundamental drivers of conflict.

This is how we end up with a sort of meta-COIN, a universalization of vague and arguably specious tactical principles to the level of foreign policy: a global emphasis on development, the effective provision of goods and services, and capacity-building (both in governance and security); counterterrorism in the very most literal interpretation of the definition, emphasizing both the indirect actions that influence and render global and regional environments inhospitable to terrorist networks (economic development, conflict resolution, civil governance capacity-building) plus direct action against terrorist networks... but the easier and cheaper version: direct action by partner security forces (military assistance, foreign internal defense, security force capacity-building). And then you've still got the ultimate trump card in reserve: direct military action by U.S. forces, which Kilcullen insists is basically the worst-case, avoid-at-almost-all-costs scenario.

What I'm getting at is this: while advocates of escalation in Afghanistan may have used the troop presence ---> security ---> popular confidence ---> actionable intelligence ---> direct action against terrorists rationale to help justify their preferred operational approach, it's hard to imagine they were being totally serious and/or honest with themselves. (How were COIN ops in Afghanistan going to provide actionable intelligence about terrorist networks and individuals that reside almost exclusively in Pakistan?) Most COIN advocates are (were?) also believers in the bigger idea, the state-strength paradigm; they saw COIN as a means of sanctuary denial, as an extended process through which Afghan society would be inoculated against the threat of insurgency and by extension, al Qaeda influence. And state-strengthening is/was the answer for Pakistan, and Yemen, and Somalia, and... everywhere, right?

What's wrong with all that, then? Well, the whole concept might be totally bogus, for one thing. We don't really understand the root causes of civil violence, even at the sub-national level. There are a whole bunch of different kinds of civil violence, and localized insurgency with a terrorist component is very, very different to international terrorism (even if you want to identify that as a component of "global insurgency" (pdf)). We can't say why, for example, one country's contentious social issues spark civil war, while those of a similar country do not, at least not with any real certainty... and that question isn't half as complicated as why some disaffected individuals choose to engage in international terrorism against a distant adversary, one that is almost certainly unconnected in any clear way to whatever tangible, material grievances that individual may have, while others join a political party or a local insurgency or just do a lot of drugs. And that's something we ought to think about before we decide to re-order our foreign and security policy around the imperatives of global development, state strength, and capacity-building. There may be plenty of good reasons to prioritize state strength, but it's not totally clear that counterterrorism justifies it.

To parrot Spencer: Let's consider this, shall we?

8 comments:

  1. Oh. My. God. Betty.

    No, No, No.

    First off, the killing of Bin Laden isn't a confirmation of CT techniques. It took ten bloody years, was extremely high risk, and probably couldn't be repeated if we sought to do so again. Imagine if we were dealing with a HVT who couldn't attract the same levels of funding and time as Osama. I read that US helicopters used a rare strike path: the channel left open to US aid flights for the Pakistani flood relief operation.

    Second off, it isn't a renunciation of COIN in terms of capturing/killing HVTs, largely because American troops were not conducting COIN in freaking Pakistan.

    If Bin Laden was killed or captured in Afghanistan due to CT efforts, I'd most likely agree, but we are comparing two very separate issues here.

    Furthermore, the following is still more or less true:
    "Proposed by Dave Kilcullen and others, counterinsurgency theorists argue that theirs is the truer path to robust counterterrorism, the pursuit of which is the only reason counterinsurgency is germane in the U.S. context. Counterterrorism requires intelligence on terrorists and their environs. Intelligence requires local knowledge. Local knowledge requires the protection of vulnerable populations from reprisal, plus some incentive -- money, justice, political access, whatever -- to make cooperation with the anti-insurgent faction sensible. If you don't provide those resources and meet people where they are, then your counterterrorism hunt is fruitless."

    A CT effort against Bin Laden in Afghanistan would have been met with a host of different factors. Bin Laden's outlook and actions were probably very different from how he would have behaved while hiding in Afghanistan against a pure CT effort considering how deep into Pakistan he was.

    "Some COIN advocates sold their operational approach by saying that it involves the collection of intelligence that enables accomplishment of the primary national objective, which is the destruction of al-Qaeda and its associated movements."

    Who posited that COIN in Afghanistan would lead to the destruction of Al Qaeda deep in Pakistan?

    "As such, we should reconsider whether the massive expense, reduced operational readiness, and other sacrifices being made to enable that COIN campaign are actually essential to accomplishment of the fundamental goal; i.e. is the war in Afghanistan really accomplishing anything vis-a-vis the president's stated objective, which is, lest we forget, to "disrupt, dismantle, and defeat al Qaeda"?"

    YES! We should, BUT NOT BECAUSE OF THE KILLING OF OSAMA BY CT ELEMENTS!

    I don't feel like we are in too much disagreement, but let us not contort and bend at great lengths to draw shallow conclusions on COIN's on-the-ground efficacy because CT finally had a stroke of success after 10 years.

    -Deus Ex

    ReplyDelete
  2. Surely this is a problem with my explanatory powers and not your reading comprehension, but you seem to have missed the point entirely.

    For one thing, I believe I made the point that the conceptual grouping "CT techniques" isn't a particularly meaningful one, as basically everything that's done in Af-Pak constitutes CT. But:

    First off, the killing of Bin Laden isn't a confirmation of CT techniques.

    I'm not sure that's what I said, but, well, yes, it is. Bin Laden is dead. Whatever we were doing worked, even if only this once. I'm not sure 100K troops in Afghanistan played a meaningful role in that.

    Second off, it isn't a renunciation of COIN in terms of capturing/killing HVTs, largely because American troops were not conducting COIN in freaking Pakistan.

    I don't know what this means, but again, I believe I made the point that it's difficult to imagine Afghan COIN meaningfully contributing to counterterrorism efforts that are focused on a group based and resident almost entirely in Pakistan.

    Who posited that COIN in Afghanistan would lead to the destruction of Al Qaeda deep in Pakistan?

    As Spencer and I both indicated, a number of advocates of escalation threw in the tack-on bonus, at the very least, that a large troop presence would result in more useful intelligence with which to target terrorists and their networks. I can likely dig up quotes if you like (and Spencer probably should have in order to make his point more clearly), but we're not making this up.

    YES! We should, BUT NOT BECAUSE OF THE KILLING OF OSAMA BY CT ELEMENTS!

    Again, this is a point that I made pretty clearly: the killing of bin Laden isn't the reason to re-examine the mission -- the utter stupidity of the counterterrorism rationale being used to justify the war is reason enough.

    I don't feel like we are in too much disagreement, but let us not contort and bend at great lengths to draw shallow conclusions on COIN's on-the-ground efficacy because CT finally had a stroke of success after 10 years.

    The point I'm making isn't about the "on-the-ground efficacy" of COIN qua COIN, but rather about the silly argument that COIN is the most effective way to do CT in a neighboring country.

    ReplyDelete
  3. Except that Bin Laden was tracked down and killed in an area of high state strength. The relative stability of the area was critical to the surveillance effort that developed the intelligence on which the raid was based. If he had been in a cave in the FATA, the CIA couldn't have established a safe-house from which to work.

    Also, your claims about the literature on the causes of civil violence are vastly overblown. There's plenty of garbage out there, but the intersection of work on the micro-dynamics with macro indicators have been moving towards greater clarity over the last decade.

    It's also a little rich to cite an article by Nick Sambanis about the weakness of empirical results. He's done some good work, but the data coding in his work on peacekeeping is inexcusably bad.

    ReplyDelete
  4. MK -- I have to be honest: I'm confused by all three of your points.

    Except that Bin Laden was tracked down and killed in an area of high state strength. The relative stability of the area was critical to the surveillance effort that developed the intelligence on which the raid was based. If he had been in a cave in the FATA, the CIA couldn't have established a safe-house from which to work.

    While this is surely true, I'm not sure I understand what you mean to suggest by it. Yes, bin Laden was killed in an area of high state strength. But the fact that he was LIVING in an area of high state strength would seem on its very own to suggest a reconsideration of our ideas about where terrorists live and thrive. Is it easier for a strong state to prosecute targets that threaten its security than it is for a weak state to do so? Of course. But Pakistan is a thriving example of a state that perceives threats differently than we do; is that state more dangerous to us when it is weak or strong? The strength of the Pakistani state didn't facilitate the killing of bin Laden in any of the ways that the state-strength crowd suggests it should have, like for example through action on the part of capable Pakistani security forces; it merely allowed for a sufficient level of urban development that a few U.S. citizens could blend in there. So what?

    Also, your claims about the literature on the causes of civil violence are vastly overblown. There's plenty of garbage out there, but the intersection of work on the micro-dynamics with macro indicators have been moving towards greater clarity over the last decade.

    I'm not sure which claims you're referring to. I merely said that while we may have a decent grip on correlation, we're less good at causality. I don't deny that dedicated, detailed study of a set of circumstances in one country can lead to better understanding of conflict and violence. The point that I'm making is that there's basically no empirical support for such sweeping claims as "terrorism exists because of poverty/inequality/unemployment/state weakness." I don't think you'd disagree with this statement.

    It's also a little rich to cite an article by Nick Sambanis about the weakness of empirical results. He's done some good work, but the data coding in his work on peacekeeping is inexcusably bad.

    Again, we're not really talking about "the weakness of empirical results" here so much as we are the wide variety of causal explanations for violence. The Sambanis paper doesn't say that empirical results are useless, but rather that many of them don't allow for extrapolation across different conflicts because of a lack of robustness -- that is, because of a failure to consider the qualitative impact of some contextual factors. It's strange to me that you're resorting to a "consider the source" argument to contest a contention neither I nor the guy in question have really made.

    As I said at the beginning, I'm confused by your comment. I know you likely have strong feelings about the subject I've covered here, but these three points seem peripheral. What's the meat of your gripe?

    ReplyDelete
  5. Actually, they're pretty central to your argument. You seem to be castigating those who proclaim a single unified theory of how civil violence - a category considerably broader than terrorism - arises, and how it threatens US national security interests. You then argue the evidence doesn't support a single unified theory, and therefore that the argument that addressing that violence is in our national interest is specious.

    I agree that those who try to build a single unified theory of civil conflict and its implications for international security make overly broad claims that aren't well-supported. What an immersion in the literature - especially case studies and studies of microdynamics - does provide is an appreciation of how a broad array of factors can combine in multiple ways to generate civil violence. These are Mackie's INUS factors - individually necessary for a particular causal pathway to produce outcome A, in which that pathway is sufficient but unnecessary to produce A. The same factor may be present in more than one distinct causal pathway, and play a slightly different role in each one, but it doesn't mean the factor is irrelevant.

    When looking at a given case, familiarity with the variety of ways that factors can combine in causal pathways can help identify the key factors without eliding the idiosyncrasies of the case. This doesn't mean the explanations aren't valid, or that we can't build a nuanced policy around them. My problem with your invocation of Sambanis is that the requirement for quantitative robustness across disparate cases obscures these realities by demanding the identification of a single basket of factors that are positively correlated with all civil conflicts. It uses a methodology that, IMO, privileges parsimony at the expense of explanatory power. This is typical of the over-simplified linear explanations that underpin too much of our foreign and security policies, and is the basis for dismissing out of hand the consequences of a precipitous withdrawal from Afghanistan. As I've written before on this blog, explaining the logic of preventive is difficult because one can't point to a single pathway that will definitely occur, but rather an array of pathways that all lead to bad results. Each one may not have a high probability, but cumulatively their probability may be significant. I think this is the case in Afghanistan, and that, in the same way we once wouldn't have logically linked stationing US troops in Saudi Arabia with a nightclub bombing in Bali, the reverberations of an insurgent victory there are difficult to predict with any precision, but would be significant.

    ReplyDelete
  6. I appreciate you taking the time to explain, because this makes a lot more sense to me now. Whether or not this constitutes irony (and I'm not sure), you've actually helped to highlight the root of our disagreement.

    The criticism of Sambanis for favoring parsimony over explanatory power is perfectly sensible and legitimate. But when you write that this approach is flawed for "demanding the identification of a single basket of factors that are positively correlated with all civil conflicts," I think you've gone a bit too far. Perhaps I'm being too generous, but I understand the research question somewhat differently: as an effort to differentiate between the perfectly legitimate causal factors associated with civil war onset that are generalizable and those that are wholly (or at least significantly) context-dependent. (I suppose here I would say that EVERYTHING is context-dependent, but if we start from that point, then science is both impossible and pointless.) To my mind, the benefit of a study like that is to help clarify the difference between necessity and sufficiency.

    But now I'm starting to get a bit vague myself. More to the point: you deride "the over-simplified linear explanations that underpin too much of our foreign and security policies" and suggest that such generalities are the basis for most opposition to the war. I'm similarly frustrated, though -- and this is where the irony comes in -- this whole post was about the way that over-simplified, linear and specious generalities are so often used to rationalize and justify the continuation of the same war.

    It's interesting that you invoke the supposed relationship between US force posture and foreign terrorism, because I would again respond that we don't really know how troops in Saudi and bombings in Indonesia are related (beyond rhetorically). Cribbing the SECDEF's refrain about our success in predicting the next war: we have a perfect record when it comes to understanding the second- and third- and x-order effects of our policy decisions -- we've never once gotten it right. Which is really what it all comes down to: one side says "we can't know with any certainty what will happen, so we should do X!", while the other side says "we can't know with any certainty what will happen, so we should do Y!" And so it's left to all of us to use the best available data to analyze things in context and make policy recommendations as best we can... but without abandoning our agnosticism about root causes or epistemological humility or whatever else you want to call it.

    THAT was really the point of the post, not just arguing for one COA or another.

    ReplyDelete
  7. Which is really what it all comes down to: one side says "we can't know with any certainty what will happen, so we should do X!", while the other side says "we can't know with any certainty what will happen, so we should do Y!"

    Actually, this bit wasn't totally accurate. I should have continued by saying that there are SOME people who are not similarly willing to confess agnosticism and admit epistemological limitations, and THOSE are the people we should be arguing with. It's not the people who reach different conclusions that I have a problem with; it's the people who use fabricated, bullshit rationales to justify those different conclusions.

    ReplyDelete
  8. I whole-heartedly agree with your point about specious rationales, but it seems to me that while there are plenty on both sides of almost any foreign policy argument, opponents of longer-term and higher-intensity engagements/commitments tend to heap unjustified scorn on complex causality. Sound-bite (or blog-length) arguments trump real knowledge and understanding.

    But once again you're arguing that we know less than we do. Certainly there are bullshit arguments, but not all of them are. And you're overlooking how INUS factors work - it's not about distinguishing generalizable factors across all cases, but how factors come together in multiple ways. Looking at the factors individually across all cases won't reveal that, and a reliance on that approach obscures how much we've learned.

    I also disagree with the assertion that we (and here I don't just mean the US - I mean allies and IOs) never manage to successfully shape outcomes (a huge extrapolation from SECDEF's original point). That's just not true. The road is almost always bumpy and riven with missteps, but there are examples of relative success. They usually emerge from a conscious effort to engage with the complexity of a situation, and a concerted effort to balance between short-term and long-term priorities.

    ReplyDelete