Tag: formal theory

The Logic of Violence in Counterinsurgency

I have alluded to the work of Jason Lyall on the use of indiscriminate violence in counterinsurgency in the past. Briefly, Lyall’s paper (recently published in JCR) examines how the Russian army used targeted and non-targeted shelling in Chechnya through a pseudo-natural experiment. The paper is fascinating, however, I always had two major issues with it; first, Lyall claims randomization and thus indiscriminate violence through the “harass and interdiction” pattern of shelling used by Russians. With even my limited exposure to US Army protocol, it is difficult to claim that this pattern is truly random. More importantly, though, is Lyall’s always struck me as an extremely useful empirical analysis in search of a theory.

A recent working paper entitled, “The Political Economy of Counterinsurgency Violence,” seeks to fill this void by offering a simple formalization of counterinsurgency strategy. In fact, the author ask an extremely important question in the opening paragraph:

Why are counterinsurgents often so brutal toward civilians if classical counterinsurgency theory is correct in suggesting that successful counterinsurgents must win—not destroy—the hearts and minds of the population?

To understand this dynamic the author models counterinsurgency as a game with three players. First, a coalition bwtween a rebel group and their popular support within a community, and second the counterinsurgent. To achieve its goals, the counterinsurgent seeks to divide this coalition through a mixture of violence and concession, which tempts some side in the coalition to defect on its partner for short-term gains and forgo long-term goals. Formally, the game is played as a public goods game, where each player has some level of “profit” it extracts from the insurgency, which is offset by the cost of participation. Thus the counterinsurgent seeks to short-circuit the profit chain through the threat or execution of violence.

What falls out of this model is an very interesting observation about how insurgency are a function of the active micro-economies where they take place. As the author states:

The rebels’ profit from insurgency increases due to windfall and black market revenue, external aid, natural resources, taxation, remittances, looted property and labor, and the availability and attractiveness of the rebels’ sanctuaries. An increase in the rebels’ accountability to the population and a decrease rebel profit results from restrictive geography, vulnerability to the population’s disloyalty caused by the nature of the rebel group’s organization, and the presence and strength of quasi-judicial institutions with which to sanction rebels’ abusive behavior…Factors negatively and positively affecting the actors’ relative profit during insurgency ought to correlate with the government’s use of indiscriminate violence.

The model is clever, and the author’s keen attention to the work of key counterinsurgency scholars comes through in his incorporation of critical elements of insurgency in the model. What is interesting, however, is how the model does not do a good job of predicting the kind of indiscriminate violence observed by Lyall in his research. The author here uses case studies from Guatemala and Turkey to support his theory, but given the high profile of Lyall’s work it would have been much more satisfactory to get a model that explained those observations. Of course, it is not the job of a modeler to fit data, and it may simply be the case that Lyall’s natural experiment is flawed, and this model requires better data for testing; either way, the article is very engaging and I highly recommend it.

Photo: New Internationalist


Armageddon and discount rates

A longstanding plan in my circle at Columbia was to take a large number of existing game-theoretic models, apply an infinite discount rate (rendering the value of future gains and losses effectively zero), and submit our findings to Rationality and Society as an article entitled “The Impact of Belief in the End Times on Common Non-Cooperative Games.”

So I’m sad to report that anecdotal evidence suggests our findings, had we proceeded with the plan, might have lacked any empirical validity. Tim Burke:

The scene: the supermarket checkout line this afternoon. The woman ahead of me and the clerk are having an animated conversation.

Clerk: “I’ve read the Left Behind books, you know. It makes you think, it really does.”
Woman: “Yes, it’s just like Revelation now.”
Clerk: “Completely.”
Woman: “You know Our Lady of Guadalupe? Well, she’s from Mexico City too. So it makes sense that it would start there.”
Clerk: “Though I thought it wouldn’t be until 2012.”
Woman: “You have to be ready to meet Our Maker anytime. I think this is it, though.”
Clerk: “The Aztec calendar is more accurate than ours, isn’t that true.”

Woman finishes paying, walks away. As I leave the store, she’s looking over her receipt carefully and heads back into the store, looking to question something on the bill. As I head out the door, I look back and she’s energetically showing the receipt to the manager.

But I can’t help wondering.

Perhaps the woman was simply hedging her bets? She might, after all, worry about being one of those left behind….


Fuzzy Logic

This week, as part of my “Theory and Beer” meeting with doctoral students, we considered Michael Gilligan’s thought-provoking formal theory article on the International Criminal Court, published two years ago in International Organization.

Gilligan uses a formal model to “test” whether the International Criminal Court can be an effective judicial institution given that it has no police arm and, as is so clear with respect to Darfur, cannot force states to extradite indictees.

His argument in a nutshell:

“Although the institution has no enforcement power, some leaders (those with such a high probability of being deposed that they would willingly surrender to the institution rather than try to stay in office) will be punished by it [because] in those circumstances… foreign state(s) have no incentive to offer the leader asylum. [Thus], the institution may deter some atrocities at the margin.”

In other words, if atrocity-committing dictators are about to be deposed they will prefer some form of exile to the possibility of torture and execution; they will prefer exile in freedom in a third party state, but they will turn themselves over to a court if third party states do not offer them exile; and such states would have less incentive to make such offers, presumably, now that a plausible alternative – jail in the Hague – is available as a means of removing said dictator.

Gilligan’s model works flawlessly, given its assumptions, but those assumptions are not well substantiated in his piece. In particular his argument hinges on the idea that “foreign countries prefer not to have atrocity-committing dictators in power in other countries and so cannot credibly commit to deny asylum to an asylum-seeking human rights abuser” (p. 943).

I can see at least two problems with this assumption. First, I see no basis for thinking foreign countries care a whit for atrocity-committing dictators remaining in power. If they did, why would they wait until the person was about to be deposed to try to buy them off? Second, I am unconvinced that decisions to extend exile are based on the desire to remove them from power. There are many other possible explanations, the most banal of which is that many other sitting heads of government will have personal connections to any particular dictator as they are all part of the diplomatic club. Why else wouldn’t they be just as satisfied to see the dictator killed Samuel Doe style by the angry hordes?

Some interesting recent evidence that the “third-party exile” norm is well-entrenched: Malaysia has just offered exile to embattled Zimbabwean President Mugabe. Admittedly it’s not a fair test of Gilligan’s model since the ICC hasn’t indicted Mugabe for anything. But does Gilligan really think that if it had, Malaysia would not still be “welcoming Mugabe with open arms?”

If the incoming President ratifies the ICC Statute and the Prosecutor issues an indictment, this event may yet provide a helpful test of Gilligan’s model. In that case the question would be whether Malaysia is willing to extradite. My guess: without amending the ICC statute to make it an international crime to “harbor former dictators” Gilligan’s model won’t matter much even if it’s right. Anyway, until we see some empirical data, a study like this remains purely abstract and nonsensical.


© 2020 Duck of Minerva

Theme by Anders NorenUp ↑