Annette Koh

Oops. Forgot to add the reference. From the Final Report of the National Commission on the Causes of the Financial and Economic Crisis in the United States which is the best-written 600 page government report ever:

As early as the 1970s, Wall Street executives had hired quants—analysts adept in advanced mathematical theory and computers—to develop models to predict how markets or securities might change. Securitization increased the importance of this expertise. Scott Patterson, author of The Quants, told the FCIC that using models dramatically changed finance. “Wall Street is essentially floating on a sea of mathematics and computer power,” Patterson said.

The increasing dependence on mathematics let the quants create more complex products and let their managers say, and maybe even believe, that they could better manage those products’ risk. JP Morgan developed the first “Value at Risk” model (VaR), and the industry soon adopted different versions. These models purported to predict with at least 95% certainty how much a firm could lose if market prices changed. But models relied on assumptions based on limited historical data; for mortgage-backed securities, the models would turn out to be woefully inadequate. And modeling human behavior was different from the problems the quants had addressed in graduate school. “It’s not like trying to shoot a rocket to the moon where you know the law of gravity,” Emanuel Derman, a Columbia University finance professor who worked at Goldman Sachs for 17 years, told the Commission. “The way people feel about gravity on a given day isn’t going to affect the way the rocket behaves.”

You can also watch this video for a surprisingly thorough rundown of what the heck happened.