Sunday, January 16, 2005

Troop Strength Again

The Belmonst Club has a useful post on what adequate troop strength in Iraq should be. He highlights numbers I've used in the past when I've looked at the numbers and concluded we have enough troops in Iraq to win based on historical data (although I concede that local factors are more important than any historic average. We are fighting this war and not the historic average of wars).

Wrethcard posts a chart (man, I really need to figure out how to do charts on this site) that shows that for policing, security forces representing 0.22% of the population is sufficient. Light resistance may require 0.4% to 1%. A serious insurgency requires 2% or more. This is fine.

What I challenge is his chart showing that in Iraq as a whole we have 0.6% security levels. Assuming 25 million Iraqis, this percentage clearly refers to only the 150,000 American soldiers in Iraq. It is this kind of thinking that made me reject General Shinseki's famous statement that we would need several hundred thousand soldiers to police Iraq. Not that I rejected the absolute numbers given Iraq's size, but it was and remains a mistake to assume this has to mean only American troops.

American forces are not the only security forces in Iraq. We have (and I'm going purely on memory here so forgive any errors--the general idea will remain valid):

  1. 150,000 Americans.
  2. 25,000 Coalition.
  3. 105,000 Iraqis trained.
  4. 30,000 Iraqi militia (excluding Kurds)
  5. 20,000 Contract security.
This totals 330,000 security personnel.

We should not count this against the full 25 million population since the security forces above exclude Kurds which at one time anyway numbered 70,000. But since Kurdish areas are secure, let's just exclude the Kurdish population and we don't need to count the Kurdish militias.

So we have 330,000 security personnel for 20 million people. This is 1.65%. This isn't enough for security but we have to look at Iraq and not the average. We also have 25,000 personnel in Kuwait supporting the Iraq force. How do we count this? If Kuwait didn't host them, presumably they'd be in Iraq itself. If in Iraq, we'd count them as 25,000 fighting the insurgency. Why should we exclude their value just because they are in a more secure area and not subject to insurgent attacks? Indeed, since they don't need the security of Iraq-based forces, shouldn't we value them higher since they are doing the equivalent job of 25,000 plus security if they were inside Iraq? Let's call them 30,000 assuming a brigade of security would be necessary. I'm being conservative since our 20 brigades in Iraq at 5,000 per brigade works out to 100,000 in combat units and 50,000 non-brigade troops.

Now we are up to 360,000 troop equivalents in Iraq. Now we are up to 1.8%. Close but no cigar.

But in my past back-of-the-envelope calculations I had divided troops into Kurdish areas (5 million), Shia areas (15 million), and Sunni areas (5 million). Using this method I found sufficient troops although this ethnic breakdown is simplistic since ethnic/religious groups are not distinct in their own geographic areas. Let me assume roughly over half of the Shias and Kurds are in mostly "pure" regions (so I can round to even numbers). So assume that we can really subtract from 25 million total population 3 million Kurds in safe Kurdish areas and 8 million Shias in Shia regions in the south. This leaves us with a population of 14 million to measure troops strength against.

How are the troops occupying Iraq broken down? I'm going to do more assuming here for the numbers in the pacification area:

  1. 150,000 Americans.
  2. 20,000-equivalent Americans in Kuwait (I assume some support the Shia areas).
  3. 10,000 Contract security (half).
  4. 75,000 Iraqis (about 3/4 of total).
  5. 10,000 Iraqi militia ( 1/3 of total).
This totals 265,000 in the pacification area leaving 95,000 for the 8-million Shia area. This is almost 1.2%. Well above the light resistance range that characterizes the Shia area. Given the security situation, this is a high presence. At the 0.5% level we'd only need 40,000 security personnel in the Shia area.

Back to the pacification area with Sunni or mixed areas with lots of Sunnis, I'm assuming 8 million people. The 265,000 I've assigned to this area represents 3.3%. This is well above the historic average of fighting serious insurgencies. And it also gives us a reserve of forces in the Shia areas in excess of its policing requirements and from the Kurdish forces that I don't even count. Perhaps tens of thousands could be moved to the pacification area if needed. In addition, at one time we counted over 200,000 Iraqis trained and untrained. I assume that the 100,000 not trained are somewhere and if not tested should count as something. Even counting them as 1/4 effective would give us 25,000 more and if in light policing areas would do just fine.

Given the money the Sunnis have to fund their fight, their skill level, their motivation, and their jihadi allies, we might want up to 4% in the pacification area but given the lack of precision in my assumptions, that we are well above the 2% level is reassuring. The fact that the insurgency has not derailed the standing up of Iraqi security and governmental institutions is the main sign to me that we have enough, validating my reassurance.

I would like to note that I did the calculations as I made my assumptions and did not work backwards from a desired level. My assumptions may be wrong but I did the calculations honestly here. Indeed, I think I am conservative in my assumptions. If I ever run across more precise population data and troop deployment numbers I'll revisit this. But I'm blogging and not writing for publication so I'm not highly motivated to bolster my memory of past data I've read about with actual research (sorry).

We have the troops to win. Replacing US troops with Iraqis is more important than arguing about US troop levels.