- Managing Me
- Big Ideas
- Managing People
It must be something about October.
From 12.11pm to 3.56pm yesterday, the Australian Securities Exchange was unable to publish market announcements. Trade was halted for 23 companies, including steel-maker Arrium, when they submitted announcements that the ASX was not able to relay to investors. As the market closed, traders were flooded with hundreds of company announcements, leaving traders with a backlog of information to digest.
It’s not the first time in recent memory that the ASX has been felled by a technical glitch. A year ago, on October 27, the ASX had to shut down for four hours as it sought to remedy a computer meltdown.
While yesterday’s situation doesn’t appear to have caused too much damage (by the closing bell the ASX broke the 4500-point barrier for the first time since July 2011, even though volumes were low), it highlights how vulnerable global sharemarkets are to computer errors.
If the global record shows anything, it’s that in calm circumstances, such moments of confusion are manageable. Trades can be reversed, and markets generally recover at least some of their losses. But if such an error happened during a period where prices were already rapidly falling, it could trigger a proper ‘run’ on the market, with everyone scurrying to sell out before a complete crash.
In investigating yesterday’s glitch, the Australian Securities and Investments Commission joins dozens of regulators worldwide currently looking into one error or other that has recently caused share price carnage.
Here are four computer glitches that have recently bought global bourses to a partial or complete halt.
On October 3, the NASDAQ cancelled trading in Kraft after a computer error sent the snack company’s price soaring 30% higher within minutes of the opening bell. Later that day, the exchange said it would cancel all trades in Kraft above $47.82 that were executed between 9.30am and 9.31am. It called the $1.5 million worth of trades “clearly erroneous”, without elaborating.
It’s not the first time this year that glitches on the NASDAQ have hurt individual stocks: Facebook’s IPO (on May 18, 2012) sent NASDAQ’s system crashing under the weight of demand for the stock. Early trades were delayed for half an hour, and after officials at NASDAQ said the glitch was fixed, more issues with trading followed. For the first half of the day, traders were unable to verify whether their Facebook trades had gone through.
Then, on August 1, a software glitch threatened to send listed broker Knight Capital into a “near-death spiral” (according to its chief executive). The prices of about 150 stocks mysteriously skyrocketed. It was eventually revealed that Knight’s own high-frequency trading platform was responsible, due to problems with the algorithm it used. Officials cancelled all trades executed 30% above or below a stock’s opening price on that day, and Knight Capital lost $440 million.
Of course, any summary of America’s market errors has to mention Wall Street’s ‘flash crash’ of 2010. For a horrifying 20 minutes on May 6, prices on Wall Street plummeted. It was a bad day to begin with: European debt worries made prices more volatile. In response to that volatility, computerised trading systems halted for a period, which was read by other algorithms as a bidding down of stocks. The Dow Jones Industrial Average went into freefall, wiping out 998.5 points, and erasing $863 billion. The fall continued until an automatic stabiliser on the futures exchange kicked in, pausing trading for five seconds. This broke the cycle, and the market recovered. The American Securities and Exchange Commission has since expanded the use of such circuit-breakers, which now halt trading for five minutes if a stock’s price moves by 10% or more in a five-minute period.
Last Friday India’s main bourse stopped trading for 15 minutes, causing the main share index to tumble 16%, after brokerage firm Emkay Global Financial Services placed 59 wrong orders worth more than $125 million. The orders triggered a broader sell-off, which wiped almost $60 billion off the value of many of India’s largest companies.
Most people immediately assumed the mini-crash was caused by high-frequency trading, but it was in fact a result of human error. The orders were entered incorrectly, and then computer programs kicked in and magnified the error before anyone realised what was happening.
All trades were eventually cancelled, and the market recovered, closing down only 0.8% lower.
The Tokya Stock Exchange has suffered two system errors this year, the latest on August 6, when it halted derivatives trading for 95 minutes. The error was compounded when backup systems failed. Luckily, the market was quiet, and some trading was diverted to Japan’s second, smaller exchange in Osaka.
In February, during the height of Japan’s earnings season, the Tokyo exchange crashed for three and a half hours after a server failure. On December 29 a cable malfunction slowed trading for several hours.
On August 6, the same day Japan had its problems, share trading in Spain was disrupted for more than four hours on the Blosa de Madrid, the country’s main stock exchange, after a glitch hit the systems of the exchange’s operators. In this case, costs from the errors were minimal, with the markets ending higher at the end of the day.