Friday, March 25, 2011

"Social science and the Libyan adventure", Rebutted

. Friday, March 25, 2011

This post from Stephen Walt obfuscates some things:

Before France, Britain, and the United States stumbled into its current attempt to dislodge Muammar al-Qaddafi from power in Libya -- and let's not kid ourselves, that's what they are trying to do -- did anyone bother to ask what recent social science tells us about the likely results of our intervention?

I doubt it, because recent research suggests that we are likely to be disappointed by the outcome.

He then surveys some of the high points of the literature. The basic takeaway is:

We should all hope that Libya proves to be an exception to this tendency, but these various scholarly studies suggest that the probability that our intervention will yield a stable democracy is low, and that our decision to intervene has increased the likelihood of civil war. Heading off that possibility is likely to require a costly and extended international commitment, which is precisely what the people who launched this operation promised they would not do. We'll see.

What's the difference between those two claims? In the first, Walt alleges that the goal is to remove Gaddafi. In the second, he claims the goal is to establish democracy. Meanwhile, UN/NATO claims the goal of the intervention is to protect civilians, although it certainly hasn't hid its preference for a Gaddafi-free Libya. These goals are hardly the same, and some may be more achievable than others. So what is likely?*

As Walt says, those hoping for the UN/NATO intervention to provide a clear path to democracy are likely to be disappointed. The Downs/Bueno de Mesquita (2006) paper (ungated) that Walt mentions is a classic study. But that is not a stated goal of either NATO or the UN. Sarkozy, Cameron, and Obama may like to see a democracy in Libya, and they'll certainly pay some lip service to democracy over the coming months, but the primary goal is to stop the killing of civilians and the secondary goal is the removal of Gaddafi. This is evidenced by the fact that the intervention only occurred after Gaddafi started killing civilians. So while the installation of a liberal, secular democracy might be the ideal outcome, the coalition leaders may be willing to accept something less than that.

I haven't carefully read the study by Downes that Walt mentions, but it seems to not quite say what Walt thinks it says. It's an analysis of "foreign-imposed regime change (FIRC)" that concludes that the likelihood of civil war onset increases following FIRC. But in Libya there is already a civil war underway, so if Gaddafi falls, it really can't be considered a "foreign-imposed" overthrow in the sense Downes appears to be describing. Indeed, the intervention is a response to the civil conflict that is already underway, and appears to be largely intended to tip the scales in favor of the rebels.

So what we should be doing is asking what the literature has to say about interventions into civil wars. I'm not super-up on this literature, but Andrew Kydd wrote a survey of it in the most recent Annual Review of Political Science. (Sorry, I don't see an ungated version.) He draws a distinction in the academic literature between conflict prevention and conflict resolution, and says that third parties can affect civil conflicts through material incentives (including military intervention) and information provision. At this point I think we can agree that we're past the "prevention" possibility, and the UN/NATO intervention is probably more about material support than information revelation, so let's hone in on those corners of the literature.

The first thing that intervention can do is dramatically change the costs of fighting. If this cost-increase is large enough that it will off-set any benefits from continuing to fight (Kydd says this "expand(s) the bargaining space", sorry SBD), then conflict resolution may become more likely. Centinyan (2002) finds that the threat of intervention does not affect the probability of conflict, only the terms of the bargaining process. Favretto (2009) argues that the effect is nonmonotonic: highly-biased and unbiased third-party interveners can help the warring factions reach a settlement, while weakly-biased interveners exacerbate the situation. Regan (2002) finds that biased interventions shorten the length of conflict, by altering the costs associated with continuing to fight. Collier, Hoeffler, and Soderbom (2004) find that rebel-biased interventions shorten the duration of civil conflicts. Thyne (2009) finds that UN interventions likely do not lengthen the duration of conflict. Gent (2008) argues that interventions are not about the duration of conflict, but about the outcome of it, and finds that third-party interveners do effect these outcomes.

With the caveat that this is bit out of my realm of expertise, what does this literature sum to? Third-party interventions are often able to successfully shorten the duration of the civil conflict, and to alter the outcome of it. These effects are most pronounced when the interveners are biased in favor of one of the sides in the conflict. So to the extent that the UN/NATO intervention in Libya is interested in both shortening the conflict and advantaging the rebels seeking overthrow Gaddafi, there is reason to be optimistic that they'll be successful. To the extent that the UN/NATO interveners are primarily concerned with installing a democratic regime after the conflict is over, there is less cause for optimism.

In other words, Walt is only correct about his view of the likely outcomes if he's correct that the interveners are primarily concerned with promoting democracy, as opposed to other goals. But I'm not sure that's a reasonable assumption, especially coming from an unreconstructed realist, since normative concerns like democracy-promotion are assumed to be irrelevant for states' foreign policies.

*Of course, the preferences of the U.S., U.K., and France may not be same, and other NATO and UN members might have other interests as well. UNC's Stephen Gent has explored this relationship in his research. See here, e.g.

UPDATE: This post by Saideman, and this one by Carvin, say a lot of good stuff about politics of the UN/NATO mandate and command structure.


"Social science and the Libyan adventure", Rebutted




Add to Technorati Favorites