Fighting a Smarter War in Afghanistan Soldiers go home, but their knowledge doesn’t have to.

No substantial business sends its sales force out to sell a product without supplying them with market research. But we are doing just that to our troops in Afghanistan. We’ve spent an estimated $173 billion in fiscal year 2009 selling a product to Afghans—cooperation with their government—without much idea why some people buy it and others don’t.

On the platoon and company level, where American troops conduct ground-level counterinsurgency (COIN) in the Afghan Pashtun belt, we’re fighting a good war. During five embeds with the Army from 2007 to last month, I’ve seen lieutenants and captains survey their area of operations, collecting information on the economy and patterns of work and travel. They regularly sit down with local elders to collaborate on development and security measures.

This approach has long been recommended by COIN experts and the American COIN field manual as the way to fight a “population-centric” war—as opposed to an “enemy-centric” or “terrain-centric” war. Our task in Afghanistan is to gain the cooperation of the population; without tacit support from the people, the insurgency will wither away.

The problem is that valuable data are collected, but then aren’t analyzed, or not at the level where the rubber meets the road. What’s more, experienced soldiers leave. So most of our soldiers are operating with bare guesses about where the leverage points are in their local populations.

In many areas it’s obvious that a certain tribe’s territory or a certain district is more dangerous than others nearby. Or desertion is a bigger problem in some units of the Afghan National Army than others. But the “whys” are lacking.

Maybe people who live upstream in the irrigation system are more tribally cohesive than those dependent on them downstream. Maybe people who grow one type of crop have different characteristics than those who grow another. Maybe soldiers with certain backgrounds tend to desert. I’ve never seen an attempt to test such hypotheses.

West Point cadets “are required to take a course in Probability and Statistics,” according to Col. Tim Trainor, head of the academy’s Systems Engineering department. But from what I’ve seen, they don’t seem to be asked to use those skills in theater.

The problem goes beyond the level of districts or provinces. Commanders have come and gone. Some of their notes about local leaders, genealogies, economics, politics and culture have made their way up and down the chain of command, and some haven’t. There is no general Web site on Afghanistan that incorporates intel reports, commanders’ notes and population surveys for our men and women on the ground.

There are extenuating circumstances. Data are rare in Afghanistan. There hasn’t been a census since the 1960s, and the rural population is around 89% illiterate. Moreover, the Army’s quantitative experts were until recently concentrating on Iraq.

The good news is that more sophisticated methods are now being introduced in Afghanistan. Col. Pamela J. Hoyt heads the first team tasked with analyzing data in Afghanistan for the generals who set policy.

“What we have found, as you state, is that data is not in one repository with easy access,” Col. Hoyt wrote to me in a Dec. 15 email. She’s developing a database using previous surveys as well as “a model to evaluate if the Afghan National Army can achieve their growth objective given historical recruiting, attrition, and re-contracting rates, and increased recruiting levels.”

It seems odd that this model would follow, rather than precede, this fall’s announcements by Gen. Stanley McChrystal about the projected growth of the Afghan National Security Forces. But data management gaps permeate the Afghan war.

For example, everyone agrees that developing the Afghan National Police is crucial. Yet in Zabul Province this November, I learned from interviews that 80% of the police in the biggest town, Shajoy, don’t speak the local Pashtu language; none are from the province. Unsurprisingly, the police are hated by locals and Shajoy is racked by suicide attacks. How would you like to be stopped by a patrolman who didn’t speak English?

Not all of the police in the Pashtun belt are non-Pashtun. In Khost, I met local residents in the force. But it seems that no one on the American or Afghan side is coordinating the deployment of police to assign Pashtu speakers to Pashtun provinces. Our troops deserve a little more help than this.

Comments are closed.