On October 25, 1962, a bear tried to climb the perimeter fence at the Duluth Sector Direction Center, a sensitive U.S. military installation in Minnesota, setting off an alarm during a DEFCON 3 alert. The alarm signal was connected to a mis-wired klaxon at Volk Field in Wisconsin, and the blaring klaxon led to an immediate order to launch aircraft. Pilots of nuclear equipped F-106A’s were taxiing down the runway to respond to the start of nuclear war when the error was discovered. A car flashing its lights raced from the command post to the tarmac and
stopped the jets.
This near brush with nuclear catastrophe, brought on by a single foraging bear, is an example of what sociologist
Charles Perrow calls a “normal accident.” These frightening incidents are “normal” not because they happen often, but because they are almost certain to occur in any tightly connected complex system.
Today, our highly wired global financial markets are just such as system. And in recent years, aggressive traders have repeatedly played the role of the hungry bear, setting off potential disaster after potential disaster through a combination of human blunders and network failures.
The recent rogues’ gallery of accidents is impressive. It includes what traders are calling the “Flash Freeze,” when the NASDAQ was shut down for three hours on August 22. Just three days before, a programming error in an internal system that Goldman used to generate option orders malfunctioned, causing stock options with ticker symbols beginning with H through L to trade at just $1. On August 1, 2012, a similar glitch
cost Knight Capital $390 million. On May 6, 2010, a Flash Crash triggered a momentary decline of 998.5 points in the Dow. The market recovered in a few hours.
Further back, in a harbinger of what was to come in the financial crisis of 2008, we can look to the 1998 failure of Long Term Capital Management. The hedge fund had made a bad bet on derivatives tied to the Russian market, and its collapse threatened to cause a chain reaction throughout the world financial system. The traders at Long Term Capital were no rookies. Among their advisors were Myron Scholes and Robert Merton, 1997 winners of the Nobel Prize in Economics for their path-breaking work on a "
new method to determine the value of derivatives."
Normal accidents, like these, occur because two or more independent failures happen and interact in unpredictable ways. After studying calamities such as the Three Mile Island meltdown, explosions at chemical plants, and ships colliding in the open sea, Perrow observed that safety mechanisms put in place to make the systems safer in fact frequently trigger the final failure.
August 22 Flash Freeze is a textbook example. Arca, a pan European trading system, experienced problems connecting to the NASDAQ stock exchange (independent failure one). In attempting to reconnect, numerous orders were sent that flooded the NASDAQ systems causing them to breakdown (independent failure 2). To cope with the breakdown, backup systems were brought online. A flaw in the backup safety system software forced the shutdown (independent failure 3).
In his book
, Perrow stresses the role that human error and mismanagement play in these scenarios. The important lesson: failures in complex systems are caused not only by the hardware and software problems but by people and their motivations. Normal Accidents
Now think about complex trading systems and traders trying to earn lots of money. In these environments, it is not too surprising that one trader at JP Morgan Chase can rack up $6 billion in trading losses while the company’s CEO, Jamie Dimon, thinks everything is
under control. Or that MF Global, run by Jon Corzine the former CEO of Goldman Sachs and Governor of New Jersey, files for bankruptcy after losing $1.6 billion, $700 million of which was illegally transferred from customer accounts.
Perrow had a fairly simple solution for the problem. High-risk systems, such as nuclear power plants, should be built only as a last resort.
That solution won’t work for financial markets. We need currency hedges, futures markets, and derivatives to keep our economic systems functioning. But we also have to realize tweaking the current system will not fix the problem. Most of the supposedly strong cures implemented by legislators to date, such as prohibiting bank holding companies from proprietary trading, are inadequate as well.
So how do we make our markets less danger prone? A good place to start would be to reduce the excessive trading volumes that lie at the root of accidents like the Flash Freeze, Flash Cash, and Goldman debacle. There is no valid reason for high frequency trading to make up more than 50 percent of all stock trades, and there is no pressing need for some $4 trillion in daily foreign currency transactions. A
Tobin tax on transactions, first suggested by Noble laureate James Tobin in 1972, of as little as 0.1 percent, would significantly reduce these volumes. Smaller transaction volumes would reduce the size of accidents and possibly their frequency.
A Tobin Tax would make markets less liquid. The players in financial markets would predict doom and claim that world growth would be stunted. And it would not guarantee complete safety.
But tinkering with the current system and looking for easy ways out, as we are now, is bound to fail. We’re in danger of letting normal accidents in the financial system become all too normal.