Posted by John Bates
The age of electronic execution has brought with it a niggling problem - fat fingered trading. A fast-paced, stressful trading environment creates an ideal incubator for hatching mistakes. They are often simple mistakes and involve pressing the wrong key on the keyboard, possibly at the wrong time and/or even for the wrong thing.
Anecdotal evidence would have us believe that fat fingered trading is rife. Real-life fingers pushing the wrong buttons include incidents as recent as January when human error caused the Canadian dollar to slump in Asian trading hours. The U.S. dollar shot up from around C$0.99 to over C$1.0030 against the Canadian dollar before immediately dropping back down, Reuters reported. The spike in the rate had little long-term impact on the market as a whole. These sorts of mini-flash crashes happen on a regular basis in many instruments.
Higher profile fat fingers include:
· In September 2006 a Bank of America trader’s keyboard was set up to execute an order when a rugby ball landed on it and executed the $50 million trade ahead of schedule.
· In June 2005 a Mizuho Securities Trader sold 610,000 shares at 1 yen instead of 1 share at 610,000 yen at a loss of approximately $225 million.
· In October 2002 a Bear Sterns trader caused a 100 point drop in the Dow Jones Index after entering a 4 billion share sell order rather than 4 million.
· In May 2001 a Lehman Brothers dealer in London wiped £30 billion off the FTSE when he inadvertently keyed in £300 million for a trade instead of £3 million, causing a 120-point drop in the FTSE 100.
Human error is part of being human. But what about algorithmic error? Algos, created by humans, can also have fat finger days. Last year the New York Stock Exchange fined Credit Suisse Securities $150,000 for failing to control an algorithm that went haywire in 2007 flooding the exchange trading system with hundreds of thousands of erroneous orders.
Honest mistakes are one thing, but there are also an increasing number of incidents of rogue traders, fraud and greed-gone-wrong. One famous rogue was the 2009 so-called drunken trader - a broker at PVM Futures who clocked up a $10 million loss from trading while intoxicated. A much larger and more grandiose deception occurred in 2008 when Jérôme Kerviel was discovered to have hidden losses valued at approximately €4.9 billion at Société Générale. And a commodities trader at MFG lost $141.5 million in 2007 on a big short position in wheat futures because his management had turned off the trading limit controls. They claimed that the controls “slowed things down” in a classic greed-gone-wrong story.
Looking at this compendium of fat finger or algorithmic errors and fraudulent or rogue trading, I think we have been lucky so far that their impact has not been more serious. The scary thing is that things could have gone a lot more wrong. The flash crash illustrated how quickly things can move and how inter-related the markets are. In 2010 we have seen incidents in equities, futures, FX and oil markets. A cross-asset “splash crash” that cascades across multiple markets is theoretically possible – whether it’s accidental or even premeditated.
Fears of algorithmic terrorism, where a well-funded criminal or terrorist organization could find a way to cause a major market crisis, are not unfounded. This type of scenario could cause chaos for civilization and profit for the bad guys and must constitute a matter of national security.
So what can be done? Better real-time monitoring and market surveillance, real-time risk and internal policing – by trading firms, trading venues and regulators. The markets should be free – but protected. Real-time visibility as to what is going on and real-time response to make course corrections when needed is crucial. Discovering or predicting problems and then being able to take immediate corrective action may help to save the world.