I present "The Story of the CDO Market Meltdown: An Empirical Analysis", an incredible senior thesis written by a friend of mine, AK (full citation in the actual paper itself). It was recently cited in the Wall Street Journal (http://blogs.wsj.com/deals/2010/03/15/michael-lewiss-the-big-short-read-the-harvard-thesis-instead/) and Michael Lewis, author of "Liar's Poker" and "Moneyball", has said it is "more interesting than any single piece of Wall Street research on the subject" and cites it in the acknowledgments of his new book, "The Big Short".
However you think about the analysis (AK makes very intelligent, simple and defensible points, most of which I agree with and a few of which I'd be interested in learning more about before coming to a conclusion), the detailed empirics of this are unmatched by anything else I've read in the news, editorial pages or blogosphere on this crisis. AK compiled the extensive data by hand, because it didn't exist.
It's very, very wonkish, so you may have some trouble understanding it if you don't know how investment banks work. It may be worth a read either way, however, because you'll pick up pieces as you go, and it illustrates the incredible complexity that characterized the system (and thus allowed it to swell so far).
For those who want the "Cliff's Notes", I've posted the table of her conclusions (table 16 in the actual pdf) here: http://tfideas.blogspot.com/2010/03/above-tables-are-from-my-friends.html