Three-toed good, Peebles bad

Thanks to a commenter, I've found Cosma Shalizi's review of Gillian Tett's Fool's Gold, which includes the following:

In short, the group at Morgan decided to change the way a basic capitalist institution worked, on the basis of abstract ideological principles, without any concern for its real-world effects, or the hard-won experience embodied in the social order they had inherited. (No doubt it helped that they all considered each other super-smart.) In conjunction with their peers at other major banks and financial institutions, they created these instruments, and then engaged in a series of highly successful lobbying efforts to ensure that they remained unregulated, and indeed to make sure that they didn't even trade on organized exchanges, so as to keep the banks' customers from seeing the deals being offered on comparable securities. (Naturally, the bankers who lobbied against transparency claimed to act in the name of free-market competition.) This may not have involved any illegal, brown-paper-bags-stuffed-with-cash corruption, but it's still a striking example of how a compact, organized, and above all rich special interest group can bend the political process to its will.

It is at this point, after the victory of an elitist cadre of self-serving ideologues, that what Tett calls the "corruption" set in. The original Morgan group were rationalistic social engineers filled with hubris, but even they realized there were limits on the trick they devised. To pull it off, the bank had to estimate the risk not just of any one loan in the portfolio defaulting, but of groups of them doing so at once, which meant estimating the correlations among loan defaults. In the nature of things, this is much harder than just estimating the risk of a single loan, and in many cases, like defaults on sub-prime mortgages, the data just wasn't there to support any sensible kind of estimation. (Cf..) The Morgan team realized this, and so did only a few mortgage deals. The rest of the industry was not so scrupulous: some of them thought they had a way of letting the market figure out the correlations, without anyone in the market having any information about the underlying economic entities, while the ratings agencies, for their part, used correlations set not so much ex ante as ex ano. The result, naturally, was an orgy of leveraged risk-taking such as had not been seen since the banks and financial markets were regulated in the first place. I say "naturally" because banks compete for investors' money, and they do so by offering returns. Leverage and gambling increase your returns, at least when your bets pay off, so the leveraged, aggressive banks gain at the expense of their more prudent competitors, who can't easily prove that they're being prudent, rather than timid and incompetent. (Tett makes it clear that Morgan suffered from its comparative restraint, and, had the bubble lasted just a bit longer, would probably have been forced to join in by its shareholders.) So leverage and risk-taking tend to ratchet up, until things go bust. If the aggressive banks have had the brains to diversify, then they have by that token correlated their portfolios (even were there no correlations within each portfolio), and so they will all go bust together. Which brings us to the end of 2008, more or less.

The rest here.