Fan mail to Aaron Task, Henry Blodget, John Mauldin, the Planet Money gang and Felix Salmon

The Wired story about the Gaussian copula function is fascinating -- although I still don't understand it completely, what comes through is this...
_
investment bankers used a formula that they didn't understand to create investment instruments totalling 62 trillion dollars over seven years because they could.

I understand the Gaussian copula function, and the end result is pretty simple.
.
Remember that any principle, if it is true, can be tested in its extreme and still hold a relative value. The reality is simply that the function was designed to address correlative issues between diverse factors which created known end results (defaults). The primary input factor which was utilized CDS spreads unfortunately had only 10 year history of data. in the short term, the data worked great while house prices never went down....in the longterm, well, let's just say that sometime the nuclear bomb gets dropped.
It doesn't happen all the time, but if it does, you have better be prepared to deal with the consequences. The banks, Wall St. and the country simply aren't prepared to deal with this.

since you understand it bryan, can you explain something to me?
_
lets assume that the formula used data from a longer period of time, including recessions in the housing market. Would the formula itself remain valid, or is it the case that the results of the formula could not be used to justify the creation of CDSs?
_
In other words, does the formula actually provide a valid measure of risk when a better data set is used?
_

My experience in using past trends to predict future events is always hazardous. You should at least give a probable range of error.

Dadfox has said it. [Sorry in advance. This is a long post]
-
If you apply the Gamma factor as a constant, you always end up with tragically overstated results. There is simply no correlation factor which operates as a constant. That is the crux of the whole function. Taking away this known and making it an unknown factor or at least one with a probable range of error destroys the outcome. The one simple situation where Mrs. O'Leary's cow kicks over the pail of milk and burns down half the city of Chicago would create the chaos and uncertainty which would destroy the reason for embarking on the endeavor in the first place.
-
To answer your question, pluk, the formula was never valid. It was a method of approximation taken as giving a known value. Given two simple independent systems, it is highly unlikely that one could design an appropriate correlation factor. Given a complicated system, well, you get the gist. Additionally, the formula, indeed no single formula, should have ever been permitted permit the unregulated expansion of the CDSs. It was not only the size of the expansion but the lack of structure and rigorous stress testing to see whether the market would be sustainable. The expansion itself was a logarithmic function. Ponzi scheme...comes to mind. Even if you could guarantee that all counterparties were secured and stable....the number would quickly have grown so large that it would have been impossible to effectuate. Just graph the numbers.....from $920 billion in '01 to $62 trillion in '07!!! Finally, the formula would not provide a valid measure of risk even if a better data set was used. There is simply no way to tell. The formula was designed to require the input of a single correlation component. If you expand upon the data set and suppose that CDSs were made during years for which there was growth as well as decline in housing (in general), it is highly likely that the Gamma factor would be positive at times and exhibit low values at others. The formula would simply be unworkable.
-
Let's take the simple example that they expounded upon in the article with the idea that the Gamma could define a known correlation. Of course, now we understand this to be folly. But the prime example that makes this case so apparent is a situation in which shoes are discussed:
.
"The relationship between two assets can never be captured by a single scalar quantity," Wilmott says. For instance, consider the share prices of two sneaker manufacturers: When the market for sneakers is growing, both companies do well and the correlation between them is high. But when one company gets a lot of celebrity endorsements and starts stealing market share from the other, the stock prices diverge and the correlation between them turns negative. And when the nation morphs into a land of flip-flop-wearing couch potatoes, both companies decline and the correlation becomes positive again. It's impossible to sum up such a history in one correlation number, but CDOs were invariably sold on the premise that correlation was more of a constant than a variable."
-
With regard to the formula using data from a longer period of time, there is no data. There was no data I should say, but even if one were to create such data, it might not have been useful. Recall that people regarded Greenspan as the Maestro! He could keep any recession from being too severe and engineer sustained long-term growth the likes of which the world had never seen. Further, no doubt in large part relying upon principles of Objectivism as taught by Ayn Rand in Atlas Shrugged, Greenspan had failed to see that with but one good tug on a thread a garment can come apart. Cheering on the creation and deregulation of markets such as CDOs and CDO squared and CDS, he began to like the taste of the Kool-Aid. Capitalism would necessarily encourage players to act in the best interest of the market itself. The dog would never 'bite the hand that feeds him.' We know this is not the case. Friedman talks about this in The World is Flat juxtaposing 11/9 with 9/11. The same principles capable of creating may also be used to destroy. Splitting an atom creates power, but it is the utilization that determines whether one powers a city, creates a Chernobyl or destroys Hiroshima.

My favorite line from the Wired article: "Correlation trading has spread through the psyche of the financial markets like a highly infectious thought virus."

Mauldin is below average as an investment guru. And he rarely has ideas of his own. And he is often a late arrival in recognizing specific macro themes. There are many much better sources. And his political jabs are inane.

You know the story about the broken clock being right twice a day? Enough said.