Tuesday, November 18, 2008

FAIL - New T-Shirt Design


New FAIL T-Shirt at my CafePress.com store

Ok so this is only going to be mildly amusing to 0.01% of you but I thought I'd share... epic fail to mean-variance analysis...

Speaking of epic fail... i dug this up circa 1997... it is from fool.com. They were reflecting on the 1987 EPIC fail.... and waxing poetic on why it happened... I almost LOL'd when I read the opening text... it goes a little something like this....

The concept of portfolio insurance, indeed the name itself, reflects the ardent wishes of its creators for a utopian investment vehicle capable of minimizing the pain of investment loss. The idea was simple, all one would have to do was pay something akin to an insurance “premium,” and investment downside could be forever limited. Unfortunately, as so often occurs in the realm of human affairs, reality served an unwilling partner.

This sounds suspiciously like 'credit default swap' derivatives... the kind that helped cause the epic fails we have seen lately... Check out the actual FOOL.com article from 1997... here

On a similar thought... I was reading this paper by Fischer Black (of Black Scholes Merton fame...) entitled 'How to use the holes in Black Scholes'. It is worth checking out if you are interested in this business... he goes over 10 of the unrealistic assumptions which the model is based on. What is interesting is that even when you increase the precision of these unrealistic assumptions... the model only marginally improves... The other interesting thing he talks about is that 'portfolio insurance' is also fragile to these assumptions. Specifically... he says..."The same unrealistic assumptions that led to the Black-Scholes formula are behind some versions of 'portfolio insurance.'"He defines portfolio insurance as any strategy where you sell as the price drops and buy as the price rises... this is markedly different from the value approach or... buy low sell high. What is interesting though is his statement that "... Because of mean reversion, the market rise during 1987 caused a sharper-than-usual fall in expected return. But investors didn't see this at first. They continued to buy, as their portfolio insurance strategies suggested. Eventually, though, they came to understand the effects of portolio insurance on mean reversion, partly by observing the large orders that price changes brought into the market. Around October 19, the full truth of what was happening hit investors. They saw that at existing levels of the market, the expected return was much lower than they had assumed. They sold at those levels. The market fell, and expected return rose, until equilibrium was restored." One wonders if a similar pattern will emerge in another 10 years... we seem to be on a steady crash cycle of 10 years... 1987, 1997 (Long Term Capital Management), 2008.... I believe that heavy chain dependancies based on modeling had some part to play in all of these events... I guess this just teaches us that our highly developed GUT instinct isn't so obsolete after all... I think that if you see something that looks funny... it won't hurt that much to poke it a bit...

UPDATE:
Found this paper on the origins of Portolio Hedging from the CALTECH website
The paper is entitled The Evolution of Portfolio Insurance by Leland and Rubinstein
Caltech Course page.
Here is the link to the actual pdf

No comments: