03 July 2012

March of the Penguins

Maindonald's book opens with the following:

It is easy to lie with statistics. It is hard to tell the truth without statistics.
[Andrejs Dunkels]

While I've always been a fan of quant, since it should tie decisions to the real world, too often they're manipulated to conform to an external agenda. The Barclays evil is about as brazen as it gets. But this piece isn't about Barclays (this just in: Diamond fell on his sword a few minutes ago), but credit raters. From today's NY Times comes the story of how it was that the agencies colluded with the banks. Pretty brazen.

The story closes with a couple of meaningful quotes (note to editors: the punch line comes first in newspaper stories, not the tail end).

The court filings also demonstrate a lack of methodology for analyzing the Cheyne debt. For example, in an e-mail before the deal was sold, S.& P.'s lead analyst wrote to a colleague: "I had difficulties explaining 'HOW' we got to those numbers since there is no science behind it. The documents show that the lead analyst at Moody's noted there was "no actual data backing the current model assumptions" for segments of the Cheyne deal.

"I don't want to miss one deal because of our model assumptions either. Is there any possibility of 'tweaking' the default table to get all of this so that we don't have to compromise?"

And, of course, the quant community continues to deny that they were anything but disinterested technicians. Yeah, right. Sometimes, the Emperor is just nude.

No comments: