I read the WaPo article shortly after it was released, and I agree with your central argument. However, I think your probability calculation is wrong. You say that:
----
If thesr[sic] two bureaucracies are working independently -- and this is an important if -- then the odds that a threat would go unobserved by both bureaucracies is .05*.2 = .01 = 1%. So, by adding another bureaucracy, even a less competent one, the chances of an undetected threat getting through are cut from 5% to 1%. That ain't nothing.
---
.05*2 = .1 = 10%.
Should be .05^2 = .0025 = .025%

I think anyone "in the business" will acknowledge the WaPo reporters' point about redundancy, but some mild caveats are necessary. When I was a division director in the Office of Intelligence at the Dept. of Energy I had two kinds of analysts working for me. One sort (mostly national laboratory guys) did detailed technical analysis of nuclear weapon issues, often in collaboration with (and occasionally in competition with) technical analysts at CIA. The other sort were what we called "current intelligence analysts," with general political/military as opposed to technical backgrounds, charged with reviewing both the raw intelligence reporting from the various intelligence collection agencies (CIA, NSA, DIA, NGA, etc.) plus the published ("finished") intelligence reports from those and other agencies, mining this vast stream for information relevant to our DOE policy customers, and crafting a tailored, limited-scope "Daily Intelligence Highlights" along with specific topical reports & briefings. I believe every cabinet agency as well as the military services did the same thing for their policy customers, who all had different interests. So not all analysis at the multitude of agencies is truly redundant.

On the issue of A/B teams, I've long felt that the two major agencies (CIA & DIA) needed "Red Teams" to do creative "what if" analysis, looking for the "black swans." For example, before 9/11 a Red Team might have looked at how al-Qaeda could take suicide bombing to the next level by crashing airplanes into buildings (a theme explored by Tom Clancy in one of his novels). This would require trained pilots, unlikely to be recruited from airline pilots in the Muslim world. Therefore selected terrorists would have to enroll in training schools for large transport-category jet aircraft. There are a number of these schools around the world but they are far fewer in number compared to general aviation training schools, and nearly everyone enrolling for heavy jet training has an airline or coporate affiliation, since the cost is astronomic. Anyone lacking such affiliation would be suspect. By going through this analytical process you've narrowed the scope of intelligence collection by quite a bit; it's evident where we should have been looking.

Dan,

I think you make a good point. Yet, this depends on whether or not we think of bureaucratic redundancy as a parallel process as you have noted. Moreover, it assumes only one type of error--false negatives. However agencies also identify terrorist plots and terrorists incorrectly--false negatives--that divert resources (i.e. Maher Arar http://en.wikipedia.org/wiki/Maher_Arar ). Given a "fixed" budget and resources (I use the term fixed very loosely), having redundant agencies with high false positive rates may negate the reduction the added agencies bring in lower the false negatives. For an excellent discussion of redundancy and Type I and II errors see "Understanding the Challenger Disaster: Organizational Structure and the Design of Reliable Systems ." by CFL Heimann in the American Political Science Review (1993).

Suppose agency A has 1 piece of the puzzle, agency B has another piece, agency C etc. etc.
Bureaucrats are not known for sharing information with competitors.

My comment was apparently deleted, yet there is a spam comment... that's demoralizing.

Wrong thread, my bad!

It is clear the Post is not as concerned about reduction of the size of our intelligence community - they just want it all done by the elite and high paid feds - not contractors. Therefore, the argument is somewhat moot. I had hoped that this would lead to an overall reduction in government and TS holders. Nevertheless, I think now we are looking at the justification of hiring 200,000(+) more feds. Before Bush, there were roughly 3 million GS feds and military weighing down this country. After Bush's dramatic increase of feds and especially Obama's we may end up with over 4 million in very short order. We haven't had the financial resources to pay for the 3 mil, where are we going to come up with the resources to fund another mil? Think about this, we have one fed with a TS for just over 4 feds that don't. With a government promoting Transparency, why do we need so many TS holders? What is so big that they are hiding?

This is not an argument of feds vs. contractors. This is an argument of reducing both. We wont lose any coverage if we consolidate contracts, eliminate overlap and redundancy; look at the justifications behind requiring a TS for each job and eliminate positions where a TS is marginal at best. Do a better job at declassifying documents prior to delivering them to personnel. Bottom line, think before acting. It was the model at one point. Now it just seems there is a pervasive laziness in government. Better to blanket the entire work force with TS's "just in case", rather than doing the job.

If we follow the premise that we need to make these contracted jobs "inherently governmental" jobs, no matter what, the government will collapse due to massive overspending. Will Obama play his violin while his Rome burns?

Redundancy in the US intel community is not necessarily a good thing either. Take for example, the Office of the Director of National Intelligence. As one of the Bush children, it seems to be an agency in search of a mission.

Redundancy in the US intel community is not necessarily a good thing either. Take for example, the Office of the Director of National Intelligence. As one of the Bush children, it seems to be an agency in search of a mission. And unfortunately, it reminds me of a bureaucratic reform undertaken by the Los Angeles Unified School District (another overblown bureaucracy). In order to make things more "responsive" they added district offices, and all they succeeded in doing was to create yet another superfluous layer of bureaucracy.

Part of the problem is our federalist system. Part is interest group politics in the bureaucracy and Congressional committees. But the problems that the superusers have are all too real. The underwear bomber was merely a new intel failure.

Not only Red Teams need to be encouraged, but cross pollination between agencies needs to be created at every level of the bureaucracy. And the contractors in intel need to be eliminated as quickly as possible. They need to be replaced by real government employees. Privatization? pppfffffffffft!

While I agree that there are probably too many people doing to much of the same thing within the intelligence community, the authors of this piece have defined redundancy rather broadly. By their definition one might come to the conclusion that the U.S. House of Representatives and U.S. Senate are redundant. They both do the same thing, pass (or obstruct passage of) laws and conduct oversight, for the same ultimate consumer, the U.S. citizen.

Well, it seems that when you sing the right tune your post gets quoted by the ODNI ;)
Nicely done!
(http://www.dni.gov/news_articles.htm)