The Algorithmic Monsters Threatening The Global Financial Sy
Source: David Leinweber
Scott Patterson’s new book “Dark Pools” has some remarkable quotes. “We feel we created a monster,” says one of the creators of Island, the pioneering electronic securities exchange. I read that Ghengis Kahn had 12,000 children; in the world of electronic markets, Island may have more offspring. Earlier systems had humans in the loop somewhere, to revise or approve quotes and trades. Island was on full auto all the time. Its DNA can be found in electronic markets operated by retail brokerages like TD Ameritrade and eTrade to bigger fish like Goldman Sachs, Morgan Stanley and Credit Suisse (and for a long time, its code as well).
Haim Bodek was a senior wizard at Hull Trading in Chicago. A true Toscanini of electronic trading, he was a good part of the reason why Hull was so profitable leading Goldman to buy them for $600B in 1999. As electronic markets evolved , Patterson writes, it had become “a market so chaotic, complex and treacherous that it could cause a brilliant trader such as Haim Bodek to conclude that the entire system was rigged.”
Even long-time participants in electronic markets will learn a lot from this book. The dark secrets of the Zero Plus “never lose” strategy. The secret paths to the front of the line in what used to be the limit order books, which had acquired Dungeons and Dragons-like complexity.
Dumb Money
That group of long-time participants include me. In the late ’80s and early ’90s, we thought the artificial intelligence rules in the buy-side trading tools we developed made them clever. In the world Patterson describes, we were the “dumb money,” with a whole chapter (14) devoted to how institutional whale traders, like Fidelity, Vanguard, Blackrock, State Street, T. Rowe Price, Legg Mason and the buy side using these “clever” trading rules, may have been efficient, but they were grist for a meaner, faster bunch of AI bandits.
To the “dumb money” charge, I’m inclined to plead “guilty with an explanation.” Our time horizons were months, quarters and years, and they worked pretty well over all. But on scales of seconds, the mean AI bandits were grabbing lots of investor cash, pennies at a time. Some of these ‘bots made money ever day for years. They were eating the lunch of market makers and specialists, done in by decimalization and national trading rules.
“Dark Pools” fills in lots of details, and explains why ideas like call markets and “Yellow Flag” trading make sense if we want to make our markets safer and more stable.
Bonus Apocalyptic Fear
There is reason to worry about this on a large scale. Regulators are still hashing out the electronic market structure of the Swap Execution Facility required by Dodd-Frank. The swaps markets have capitalization/notional values estimated at more than 10 times the U.S. stock market, over $200 trillion. They include what are now known as “toxic assets” as well as the securities that build the global financial system.
Here is the killer and current apocalyptic quote, from John Bates, founder of a trading systems firm (with some dumb money users):
“Fears of algorithmic terrorism … are not unfounded. This type of scenario could cause chaos for civilization.”
Understanding how markets happened in “Dark Pools” is on the requirements list if we stand any chance of getting this right.
| }
|