Friday, August 7, 2009

Measuring Metrics

The Obama Administration is behind in developing "metrics" of success for the mission in Afghanistan. They apparently want to get it right. I am not surprised as there is no consensus among experts on counter-insurgency about how to measure success. The article cited here criticizes my favorite metric--the rate at which the indigenous population gives actionable intelligence to those doing COIN. It did not perform too well, as violence was increasing at the time that the "rat rate" was trending favorably.

However, to be clear, violence, as I think I have said before, is a lousy indicator of anything. It does matter politically here and there, but violence can increase or decrease for a variety of reasons, many unrelated to progress on the ground.

I considered last year to apply for one of the mega-grants the US Dept of Defense was setting up to do exactly this--to figure out how to measure progress (or regress) in Afghanistan. I decided not to apply for several reasons: it would distort my research agenda for years; the probabilty of success was low for a grant that would require a lot of work; and it was far enough way from what I really know well that I would have a hard time designing the proposal. And now, as it turns out, the government is spending a great deal of effort on this.

Two other takeaways from this article--this administration does not mind learning from the previous one, and would rather be a bit late and get something right.

No comments: