Nassim “Fooled by Randomness” Taleb must have nodded knowingly at hedge fund Amaranth’s blow-up this week. Here is the key para from a recent WSJ story on how it happened, despite supposedly robust risk controls:
What Amaranth’s systems hadn’t measured correctly is how much downside risk it faced and what steps would be effective to limit losses. The risk models employed by hedge funds use historic data to figure out how much money a fund can make or lose from its positions across a wide swath of the markets. But the natural-gas markets have been more volatile this year than any year since 2001, so models might not accurately predict the possibility of big moves. They also might not predict how much selling of one’s stakes to get out of a position can cause prices to fall and obliterate paper gains.
Ah, that problem again. The trouble with historical volatility — as LTCM discovered to its detriment — is that it is a crummy predictor in financial markets, just when you need it not to be.