Sunday, August 09, 2009

Being Realistic About Troop Numbers

During the Iraq campaign I constantly defended our troop strength in Iraq as adequate to winning the war. I just used back of the envelope math skills, but it generally seemed as if we had enough troops for the varied security threats in the country. Our victory would seem to give me credibiliy on the question of troop numbers in wars.

The main problems of critics of our troop strength were in assuming all areas needed the same ratio of troops to people and that all troops had to be American troops (or at least those trained to American standards) to count toward that number.

The fact is, high quality troops were needed, but lower quality forces (in training and equipment)from the level of nearly US-equivalent to barely adequate all have their role in a counter-insurgency. You don't need a SEAL to stand guard in front of a sewage plant, for example.

So in Iraq we had up to 170,00o high quality American troops plus British troops bolstered by allied and Iraqi army and police contingents of varying quality, and even fairly poor quality glorified security guards--eventually getting up to nearly 600,000. These all added up to more than the amount the numbers fetishists said we needed in just American troops. And we won with them. We won because even though we didn't immediately have 500,000 total troops (to police 25 million Iraqis) we had enough when you looked at the troop needs for the various regions (southern Shia, mixed central, Sunni Arab Anbar, and Kurdish regions.

So when experts are already calling for more troops in Afghanistan, I'm skeptical:


A member of the strategic assessment team working with the new U.S. military commander in Afghanistan says the U.S. government and its allies need to be more realistic about what is needed to win the Afghan war, and he says that may include more troops.


Our commander over there is more specific:


The assessment report also urges the United States and NATO to almost double the size of the Afghan security forces. It calls for expanding the Afghan army from 134,000 soldiers to about 240,000, and the police force from 92,000 personnel to about 160,000. Such an increase would require additional U.S. forces to conduct training and mentoring.


So we need 400,000 Afghan forces plus about 100,000 Western troops (2/3 US), for a while anyway, to win? That's not close to the standard 2% of the population (about 27 million in Afghanistan) in security forces that is often assumed as necessary for a successful pacification campaign. Clearly security needs are different across the country.

We surely need quantity, but what should they look like? Too often, the answer is that they must look like our troops--if not actually be our troops. This is simply not so. Iraq showed us this is not necessary.

Looking at the regions of Afghanistan, the capital district (3.5 million people) is fairly secure but critical so let's allocate 2% troop strength for 70,000 security forces to secure the area.

Regional Command East (7-10 million) and Regional Command South (3.2 million) total 12 million for this example. With extra troops needed to interdict the border, let's call it as needing 2.5%, or 300,000 troops.

Regional Command West (3 million) and Regional Command North (7 million) are not peaceful, but let's assume a level that I assumed for the Shia south in Iraq and call it 1%, or 100,000 troops in these areas.

That's 470,000 troops total. That actually works with the numbers our government wants from Afghanistan forces plus our forces. The problem is that Afghanistan can't possibly afford a security structure that big.

We need to consider a force mix that Afghanistan can sustain.

What if we aimed for a good national army of 100,000 bolstered by 150,000 police (really light infantry). Add 50,000 more provincial defense forces that would be like tribal levies able to fight withing a large area if given transportation assets by the central government or our forces. Add 100,000 local defense forces that are static defenders of villages and cities. And add 50,000 infrastructure guards for bridges, tunnels, roads, and electric generating sites, for example. That's 450,000 Afghan forces with 100,000 Western troops as the tip of the spear and reserve force, and enabler with combat support and combat service support functions.

If these forces in conjunction with a successful Pakistan campaign on the other side of the border manage to beat down the pro-Taliban tribes, we might then be able to reduce forces to 1% of the population throughout the country. Then we'd need perhaps 270,000 troops, including 30,000 Western troops. We could slash the central government army to 50,000 with 100,000 police in support. Infrastructure guards might go down to 10,000, and the provincial and local defense forces remain as the eyes and ears to detect enemy efforts to enter the areas.

I'm mostly pulling numbers for the various forces out of the air, but the point is we don't need uniformly high quality forces to win in Afghanistan. Once you accept that, then we need to figure out what is affordable for a decentralized Afghanistan to field to keep the country from being a sanctuary for al Qaeda again.