Being a Bayesian Strategist in Syria

I got involved in a good twitter discussion yesterday about Syria and intervention. The battle lines were the same as in the past. Shadi Hamid and I pushed an interventionist logic, the Ink Spots crew (followup posted here), Eric Martin, Dan Trombley and others pushing back.

I want to clarify a few things. There are basically three issues here.

(1) Do we have a moral imperative to act to prevent the murder of innocent civilians? We can debate this until the cows come home. This is simply a moral judgment. I subscribe to the “Spider-Man Doctrine” — with “great power comes great responsibility.” We don’t have an absolutely imperative to act, but certainly we ought to do what is in our power to ameliorate suffering. But look, if you disagree, you disgagree. I don’t see much point in focusing on this part of the debate, since really it is a matter of moral judgment, not something subject to empirical claims or counter-claims.

(2) Is intervention likely to actually ameliorate the suffering? Jason Fritz, and others, suggest no. They argue that, essentially, advocates of intervention are assuming away Syrian resistance and building in best case assumptions about the likely response of the Syrian military. Well, I can’t speak for others, but that is not where I am coming from. My point is precisely that I don’t know how the Syrian military would respond to outside intervention. And indeed, my preference is for a limited intervention, initially with airpower to probe at this particular issue.

Strategy is iterative. Part of being a good strategist is identifying those areas of uncertainty, and devising approaches that either (a) clarify/resolve those uncertainties, or (b) are resilient enough survive regardless of those uncertainties are resolved. The resilient approach in the case of Syria is a massive intervention that can defeat the Syrian regime regardless of how hard it fights. I do not support this, as my judgment is that the costs are not congruent with the stakes at this juncture. But my point is that one possible resolution of this uncertainty over the resilience of the Syrian military is that it is quite low, and that intervention can be quite inexpensive. But we can’t know that unless we adopt a policy of military pressure to probe their capacity.

I am specifically suggesting a limited intervention, not because I believe it will necessarily work — though I don’t rule it out — but because given the information lacuna under which we’re operating worst case assumptions are pushing  us into an unhealthy all or nothing choice. We need to take some action to clarify the situation.

Now, some people will scoff at this, calling it essentially, “recon by fire.” And yes, it is. But so what? Recon by fire is a perfectly valid form of military activity. Military operations often involve probing enemy defenses for weakness and to test responses. Again, I am not suggesting this as en end point, but rather part of a process of defining the problem more clearly so that we can strategize with greater confidence. I am suggesting, in short, a Bayesian, iterative approach.

Now, I am not going into this blind. The experience of Iraq and of Libya suggests to me that it is at least possible that the Syrian military is brittle and would collapse under a brief, punctuated air campaign. But I am perfectly willing to accept that this would be proven wrong. And if it is, then I think that an air campaign is still worthwhile because it would (a) degrade Syrian capabilities, and (b) punish the regime. Degradation and punishment are not pretty goals, but they are also perfectly valid intermediate responses if we end up judging that we can’t end the killings outright.

(3) The third debate is about what happens after. Does an intervention lead to massacres of Alawites? Does it cause Syria to fragment? Does it increase Iranian influence?

I don’t know. My guess is that the answers are No, No, and Probably not.

First, I don’t think we’d see a successor regime target the Alawis. There is no good strategic dispossession motive. There is no history of anti-Alawite genocidal ideology. There is some potential for retributionist violence. But on the whole, if you look at what are usually considered danger signs for mass killings, I don’t see them operating unless the Alawis decide to provoke it by launching a sustained insurgency against a new regime.

Second, Syria is, I think, a pretty well-established state. You don’t have major ethnically-defined administrative units (of the sort that doomed Yugoslavia). I just don’t see a state fragmentation scenario.

Third, Iran… I don’t know… I guess I really don’t care. I think we tend to dramatically over-inflate the Iranian threat.

But, look, is it possible that a post-Asad situation would be worse? Yes, of course. But there are no guarantees is life. The current situation is pretty appalling. It could get worse post-intervention, but could also, just as plausibly be much, much better. It is always easy to spin out scary narratives about an uncertain future. But those arguments strike me primarily as efforts to justify inaction rather than any sort of disciplined assessment of likely alternative scenarios.

Finally, Iraq… I am not sure we’re learning the right lesson from Iraq. The problem with Iraq was not what we did in 2003, or at least that was not the main problem. We made a mistake in 2003, intervening based on bad information and assumptions. But the real disaster was sticking around for eight years after in pursuit of increasingly vague goals. By the end of 2003, we’d removed Saddam, reintegrated Iraq into the international order by lifting sanctions, ensured compliance with non-proliferation goals… and all at quite low cost in American lives and treasure. Even for an unnecessary war, it was not a policy disaster. Then we stayed, and stayed, and stayed some more, hoping to cobble together some sort of political grand bargain, limit Iranian influence, and all sorts of other tertiary goals.

The lesson of Iraq ought not to be that military intervention is always bad, but rather than military occupation in service of vague, amorphous goals is bad. But a limited intervention to end mass killings is not a vague, amorphous goal. It is limited and precise, and with discipline can be accomplished without buying into an open-ended occupation of Syria.