Wednesday, October 27, 2010

Holy Principal-Agent Problems

Whether intended or not, it seems that the effort to take out the middle level commanders of the Taliban is creating some significant challenges for the Taliban leaders.  The guys (captured or killed) are replaced pretty quickly apparently, but the new generation may not be as easy to control as the previous one. 

In P-A terms (I am an amateur on p-a theory, but increasingly use it for the NATO-A-stan projects), leaders need to select agents to do their bidding and then the problem becomes one of making sure those agents do what they are supposed to do.  The first way to ensure that the agent complies with the principal's intent is to select agents who mostly closely share interests/attitudes/etc. with the principals.  One would expect that the first agents selected would be those that have the closest ties/views of the leadership.  As those agents are captured or killed, then the bosses will have to select agents who are not as close.  It would not make sense to start with agents who do not share as many views or ties, so elimination of the old does suggest that the new ones will be less control-able.

The second way to resolve the principal-agent problem is to shape the discretion of the agents.  Given the nature of insurgency and the terrain (as compared to the dynamics of multilateral counter-insurgency), it is probably quite difficult to impose restrictions (caveats) on the insurgents.

The third way is to conduct oversight.  To monitor closely the agents.  Well, in this case, it looks like the monitors may get killed by the agents:
But the new cohort increasingly decides how these beliefs are imposed on the ground: recently the Quetta Shura sent a Muslim scholar to chastise a group of youthful commanders in Paktia Province who were not following Mullah Omar’s directives; they promptly killed him.
The fourth way is to reward/punish agents when they comply or not.  The challenge for the Taliban leaders is that their capabilities in Afghanistan are mostly on the punitive side of things, and it is not clear that they pose as much of a threat to their local commanders as NATO is.

Is the decline of Taliban control over its agents and the potential fragmentation a good or bad thing?  I vote for both.  It might make the Taliban a bit more enthused about settling the conflict before they lose control as the NYT piece indicates.  But it makes it much harder for any decision to settle to be implemented.  Rather than killing one snake, we may find ourselves trying to kill a hydra.  Not good.

1 comment:

Anonymous said...

It seems you are hinting that there may be other equally unpleasant hydra-like scenarios, stemming from the potential collapse of Taliban control-namely, its disintegration/partition.

For a blog with such an echo (and an interesting map and counter-argument by Barfield, a BU anthropologist), see: http://www.nybooks.com/blogs/nyrblog/2010/oct/07/should-afghanistan-exist/