BAD MODELS, BAD PAPER

The computer models used to assess risk and benefit in the derivative markets were terribly wrong about the sub-prime mess. In the August 13 edition of the Financial Times, we learn that Goldman Sachs' programs were ridiculously off:

“We were seeing things that were 25-standard deviation moves, several days in a row,” said David Viniar, Goldman’s chief financial officer.

What's more interesting about this statment is that it's almost exactly what the Masters of the Universe said over 10 years ago. From Frank Partnoy's INFECTIOUS GREED, pp. 263:

Risk management was more art than science, and risk could not be boiled down to a single VAR [Value-At-Risk] number. LTCM's [Long Term Capital Management] VAR models has predicted that the fund's maximum daily loss would be in the tens of millions of dollars, and that it would not have collapsed in the lifetime of several billion universes, Askin Capital Management's VAR had been only about $15 million just before it collapsed. Barings' VAR models said its risk was zero.

LTCM wasn't just a bunch of greedy hacks. It has some of the finest investment minds working for it, including Robert C. Merton and Myron Scholes. These two won the 1997 Nobel Prize in Economics for coming up with a method to evaluate derivatives. If people like this can bungle derivatives trading, anyone can. What makes this worse is that derivatives are traded over the counter (OTC), with very few regulations and almost no reporting requirements. An average investor or even a seasoned stock analyst would necessarily know that a company was leveraged up the wazoo.