Technology Review has a lengthy two part series on the financial turmoil this year and I believe it supports what I have been saying since the first ripples in February. The root cause (to use Madison lingo) is not excessive Greenspan liquidity or immoral mortgage lending behavior. The root cause is Wall Street geeks using computer models to gamble with borrowed money.
The term “Quants” historically refers to back room technicians doing the math homework for the upfront traders and brokers. When sophisticated computer programming hits the back rooms, however, the lure of big money attracts top tier mathematicians using their numerical expertise to create complex number games. As in all things, however, if you don’t understand the reality in the source of the numbers on an Excel spreadsheet, then crunching can make you believe in false realities.
Part I: The Blow-Up : On Wednesday, August 8, not long after the markets closed, 200 of the smartest people on Wall Street gathered in a conference room at Four World Financial Center, the 34-story headquarters of Merrill Lynch. .... They didn't look like Masters of the Universe; they looked like members of a chess club. They were "quants,"
No one quite knew why, yet, but the market's odd behavior would turn out to be closely linked to the work of the quants. In addition to creating arcane financial products, quants have been pushing the frontiers of computer-driven trading systems, and not enough of those systems were working the way they were supposed to--or, to put it more precisely, the way they were supposed to work turned out to be counterproductive in volatile times like these.
Part II: The Blow-Up : "The products are getting an order of magnitude more complex," says Berman. "Things change slightly, and get correlated where they weren't correlated before." Or, as he put it a little less gnomically, "You can't make it without understanding it, but you can buy it."
Beneath all this beats the great hope of the quants: namely, that the financial world can be understood through math. They have tried to discover the underlying structures of financial markets, much as academics have unlocked the mysteries of the physical world. The more quants learn, however, the farther away a unified theory of finance seems. Human behavior, as manifested in the financial markets, simply resists quantification, at least for now.
Models of reality make really great games and tremendous special effects for entertainment. Algorithms for manipulating input data sets, however, are not crystal balls into the future. If the financial markets are well behaved through the end of November it should mean the big institutions have learned their lessons about blue sky microprocessor promises. Or in other words, the realization that access to lots and lots of numbers does not make you smart.
