Jumping To Conclusions

I'm obviously not an expert on the subject, and clearly Stansfield Turner is one. But does his proposal that the nation's fifteen intelligence agencies (?!?!? We have 15 intelligence agencies? Yes, we do.) be consolidated under one person really sound like the best idea? Turner suggests this as a solution to the problems with the Iraqi WMD intelligence, but I don't quite get his reasoning. The problems were the result of one or more of the following:

• Deception by the president.
  
• Cherry-picking of intelligence which supported the administration's proposals.
  
• Inherent difficulties in gaining accurate information about the inner workings of a totalitarian state.
  

In any of those cases, I don't see how Turner's proposal is supposed to help. It might make it easier to pick out a scapegoat after the fact, but is that really what our goal should be? As long as you have more than one agency, there will always be more than one assessment, so "cherry picking" will always be possible. It's not as if putting all the agencies under one roof will result in one answer; to the extent it does, we already have a process -- the National Intelligence Estimate -- for coming up with a consensus of what the intelligence shows. Moreover, to the extent it does lead to a greater consensus, is that really a good thing? Intelligence is an ambiguous business; disguising that fact by creating the illusion of consensus would be more misleading than helpful.

Certainly, it would reduce inefficiency and duplication of effort if control of all intelligence agencies were centralized. Central planning always does that. The problem is that duplication of effort isn't always a bad thing. Competition is more chaotic, but that's a strength as well as a weakness. It means that when one attempt gets it wrong, another has the opportunity to get it right. Why give that up just to give the head of central intelligence more power?