Last week, we heard about the somewhat mysterious death of a Libyan rebel commander, leading to some speculation about rebels turning against each other as they inch ever closer to (possibly) capturing Tripoli. This reminded me of a toy model I wrote down several months ago (linked here) that, frankly, I didn’t know what to do with. Ergo, I’m putting it here to see what folks think.
Here’s the basic story from the abstract:
Why do some rebel groups divert resources from fighting the government in order to fight other rebel groups before the government is defeated? I analyze resource allocation decisions in which two rebel groups divide finite resources between fighting their common enemy, the government, and fighting one another to influence the distribution of power for the power-sharing contest that follows military victory. In equilibrium, the inability of rebel groups to commit not to exploit the loser in a power-sharing contest can lead them to divert resources away from fighting the government and towards undermining each other when the government is sufficiently weak. In other words, as the prospects for defeating the government improve, rebels become more likely to work against one another, further delaying their ultimate victory.
So when (1) there’s no guarantee that rebels can trust each other to share power once the government’s toppled and (2) the chances of defeating the government start looking pretty good, rebel groups will divert some resources away from the main war effort and husband them for using against one another once victory’s achieved. That, of course, gives us an explanation for why fortunes can be both difficult to judge and pretty volatile in civil wars. Or, from the (very bare-bones) write-up of the model:
Perversely, the better the rebels expect to do against the government, the fewer resources they devote to the war in order to husband their strength for the power-sharing contest that follows victory. Neither side wishes to let the other gain a sufficient advantage, and thus they reduce their chances of victory, perhaps even prolonging the war, because of the commitment problem created by postwar control over the state apparatus.
Now, of course, the question is what to do with this thing. There’s some other work out there linking the number of groups to the duration of war, but (if I remember correctly) for different reasons, but I am, as always, open to suggestions.