[lbo-talk] origins of the housing boom

Jordan Hayes jmhayes at j-o-r-d-a-n.com
Mon Aug 29 12:14:08 PDT 2011


Ted Winslow writes:


>> [SA]
>> It was the hysterical forecast of price appreciation, not
>> the "hiding" of risk (it was never really hidden), that
>> constituted the bubble.
>
> It's not true that risk was "never really hidden."

The blowup occured because of the usual reason: if some is good, more is better. Until it's not. A number of factors came together to turn what was a rather ordinary and predictable shift of investment to a new class into an avalanche. None of them "caused" the boom; all contributed; many fed on each other. None can be ruled out. I think the effort to discard factors and search for The Ultimate Reason is doomed to fail.

Also: to the extent that you cross a cause off the list, it becomes less likely that it will be dealt with for the future.


> A very influential quant math model, David Li's "gaussian copula
> default function," widely used to determine the risk of correlated
> default on the mortgages underpinning mortgage backed CDOs, played
> a very significant role in their explosive growth.

It's true that this formula was widely used, such that when the model broke down, many had the wrong value. But I think it's naive to point at this -- as the Wired article does -- as a/the direct cause of the blowup. For instance, I think it's safe to say things would have played out much the same as it did without Li's function.

About the best thing you can say about approaches like Li's is: it works until it doesn't.

/jordan



More information about the lbo-talk mailing list