Matt Yglesias explains the crux of their disagreement:

…it seems to me that the real disagreement here is probably driven by different views about the U.S. military than by different views about Afghanistan as such. Exum believes that the Pentagon has developed powerful new operational doctrines about counterinsurgency that make it possible to achieve things via U.S. military intervention that U.S. military intervention hasn’t traditionally achieved. I read Stewart as being skeptical about that idea…

A couple of random questions/points expanding on this:

1) Do COIN supporters think that the army is well-suited to doing development work? This is redolent of a sort of universalist bias in American thinking, which assumes that the only thing keeping an individual from flourishing is oppressive circumstance. In Iraq, we had the strong form of this thesis, wherein the removal of Saddam would lead to the rise of a functional, semi-modern democracy in the heart of the middle east. After Iraq, our aspirations for Afghanistan are lower, but the belief in the mechanism remains: if only we can limit violence, good things will happen. Limiting violence is a good thing in itself, to be sure, but it doesn’t causally lead to a stable state emerging. And a stable state in control of its territory is the minimum of what we’re after.

2) Do COIN supporters even know what development is capable of? I don’t claim to know either, definitively, not being a specialist in the field. But my father worked at the UNDP his whole career, and I did a stint at an NGO after college, so I have at least some kind of appreciation of both the ambition of development practitioners as well as the crushing difficulty of actually achieving these goals. Reading the McChrystal recommendations, the following bullet point struck me:

Facilitating Afghan governance and mitigating the effects of malign actors. Success requires a stronger Afghan government that is seen by the Afghan people as working in their interests. Success does not require perfection—an improvement in governance that addresses the worst of today’s high level abuse of power, low-level corruption and bureaucratic incapacity will suffice.

This sounds like the kind of development boilerplate one might see on funding proposals and grant requests. It blithely suggests that even achieving a percentage of perfection is within our reach. I doubt that it is even in the long term. I’m virtually certain it’s impossible in the 12-18 month horizon that General McChyrstal’s report is talking about.

3) We’re not thinking about time horizons correctly at all. Development, as a formalized field, has only existed since after World War II—that’s a bit north of 60 years. During that time, the great successes have been in eradicating diseases like river blindness and polio, and in fighting famines through the introduction of improved farming techniques. Creating good governance in post-colonial tribal societies in Africa, on the other hand, has been depressingly less successful. Surely we’ve all heard that the Afghanistan project will take as much as 50 years to get right. I’d say 50 years is the lowball estimate, even if we as a country were 100% committed to the project and suffered no setbacks in the interim.

Rory Stewart’s Tory conservatism is based on experience of not just Afghanistan, but of development work in general. The COIN proponents’ optimism seems to be based on a can-do American ethos typical of its excellent armed services. While I’m an enthusiastic supporter of the type of thinking that the COIN crowd is doing, I fear that as their star has risen in Washington, their vision has become more maximalist. All problems appear to be nails for their hammer, and all previous failures have been failures of not doing it the right way. But many of these problems, especially but not exclusively in Afghanistan, are fundamentally problems of political (not economic) development, and are devilishly hard to solve by outsiders. It’s folly to assume that a fresh pair of eyes and new determination will succeed where decades of effort have already failed.