Interdum stultus opportuna loquitur...

Saturday, February 02, 2008

NonRant: Homework Time...

Note - from June 24th 2009, this blog has migrated from Blogger to a self-hosted version. Click here to go straight there.

In the normal course of events, I don't go in much for linking to other people's output; it's not because there's not a great deal of genuinely insightful stuff out there, nor is it because my ego doesn't permit me to point folks to good ideas that aren't mine.

Frankly, there is so much really good stuff on the blogosphere that mainline research houses ought to be ashamed of themselves for not doing a much better job.

Anyway - from time to time there emerges a piece that is so timely and on-point that it becomes required reading. As such, I'm going - when time permits - to establish a new (free) section over on the subscription site, which will serve as a repository for things that I think are Required Reading for anyone who has an interest in understanding the lay of the land.

It will basically take the form of a set of links with little snippets - either from the material itself, or my thoughts on it - so that viewers can get some hint as to whether they ought to bother clicking the link.

That's for the future - for now, I think that everybody ought to read a piece with the rather unwieldy title "A Little Noted Case Regarding Credit Default Swaps Could Portend Major Problems for Financial Institutions" over at OverHedged.

Pretty much everybody knows that although the 'notional' derivatives pool is now in the tens of trillions of dollars, many of those notionally 'net out' thus making the overall exposure of global finance to derivatives orders of magnitude smaller.

And there's the rub - the whole 'netting out' idea is a lot more notional than you might think.

OverHedged detail what is literally a perfect example of the problems that can arise when supposedly 'neutral' transactions suffer from unintended counterparty risk. Interestingly enough, one of the parties (the one that did not get shafted, in this case) is none other than Jerome Kerveil's playground - Societé Générale.

In other words, let's say firm A has on its book two supposedly offsetting positions, one with firm B and one with firm C. Both are valued at exactly $100 dollars. Both relate to the same 'event risk'.

Well, as OverHedged shows lucidly, something as simple as a loose contractual definition can wind up with firm A owing firm B the $100, but firm C being able to renege without breaking the law or going out of business. Firm A goes from a smartypants who managed to lay off a bet and book a profit, to someone who lost the whole enchilada.

Another sensational piece from the week is "The Great Credit Rating Scandal" by Paul Amery, which is over at  PrudentBear under the "Guest Commentary" section. It lays out, in a far less screechy fashion, the stuff that I have been saying about ratings agencies since Adam said "Hey Eve, look... those apples look ripe. Who's the snake?"

It's a longish piece, but well crafted, and if you read it you will get a handle on the ratings problem.

Also at PrudentBear (from earlier this week) is a very insightful piece called "The absent-minded doctor" by Fred Sheehan. It presents the mind-boggling idea that there are still people who think that Alan Greenspan knows his arse from his elbow... despite the fact that his only attempt to do something that wasn't paid for by taxpayers, ended in farce.


And finally, Mish over at Global Economic Analysis has a good piece on the flim-flammery in the US labour statistics - e.g., the CES Birth Death Model... another thing that I have banged on about since 2002. Mish's piece is entitled "Jobs Contract as 2007 Job Growth Revised Away".

So there you have it - about twelve pages of stuff which will augment your understanding of investment markets as they currently stand - in an era of statistical chicanery by the US government, abject failure by ratings agencies, huge counterparty risk in derivatives markets, and continued hagiography surrounding that dolt Greenspan.