Signup for our free newsletter:

Robot Trading

Written By Brian Hicks

Posted July 25, 2013

Several months ago, the computing world thought the White House was under attack and that President Obama had been injured in the chaos. As a result, the Dow Jones crashed 140 points, and around $200 billion had simply been erased from the markets.

If you’re surprised you don’t know what I’m talking about, that’s because such an attack never took place, and the President is fine. The markets did nosedive; however, everything recovered quickly enough.

algo trading machinesWhat really happened, as Business Day notes, was that the Twitter account of the Associated Press had been hacked by a sketchy group going by the moniker of the “Syrian Electronic Army.” Thus, the AP apparently published a tweet on that day noting the ‘attack.’

Promptly, the trading computers that have come to dominate trading activity on Wall Street caught the keywords and took action accordingly. Hence, the crash occurred before anyone even knew what was really going on. And those last words you read could also describe the final events in Terminator 4: Salvation, which shows Skynet’s rise and eventual takeover.

Some 70 percent of all trades today are executed by computers operating according to über-complex algorithms with minimal human oversight. All of this can be traced back to the post-Moon landing era, Business Day notes, when the U.S. government slashed science and research funding in order to continue pouring billions into the Vietnam fiasco. Several generations of newly-minted science graduates needed to feed themselves, so they took on the world of finance and turned it on its head.

Remember Medallion? The hedge fund that earned a return of 2,478.6 per cent within its first ten years of operation? That was headed by one of the most famous “quantitative analysts” ever—Jim Simons. That’s when the market took notice, and that’s when the quants started taking over Wall Street, bit by bit. It’s all down to data analyses on massive scales and with laser-sharp precision.

The key technology under focus is high-frequency trading, which basically executes massive numbers of trades within microseconds, trying to make the most of every fractional fluctuation in prices. HFT is so serious that, as Business Day points out, Spread Networks laid down a direct cable connecting New York and Chicago, boring through the Allegheny Mountains, just to cut a little more than 1/1000th of a second of the transmission time between the two exchanges. This was in 2010.

Problems Rather than Solutions

But HFT can go disastrously wrong, and has. Again in 2010, on May 6, the computers went crazy with trades, causing the online trading part of the NYSE to simply freeze up and crash between 2:30 and 3PM. The Dow lost and subsequently gained back over $1 trillion.

Amidst the chaos, Apple temporarily saw its shares shoot past $100,000, while Accenture’s shares dipped to nearly $0. It became abundantly clear that HFT—far from being a boon—actually increased volatility to dangerous levels.

And that may finally have made an impact, as Bloomberg reports currency funds that rely on computerized trading made 0.9 percent thus far in 2013, compared to 2.5 percent for those that don’t use such algorithms. More dramatically, Hong Kong’s Ortus Capital Management Ltd. lost 13.8 percent over the first half on its $1.1 billion computer-driven fund.

It seems the international volatility of the markets, combined with the uncertainty at central banks, have proven too much for these algorithms, which function better on more stable ground.

From Bloomberg:

“Central banks have become the insider traders of the currency market, which is a paradigm shift that systematic traders cannot pick up as well as fundamental traders,” Richard Olsen, the founder of Olsen Ltd., who has designed currency-trading models since 1986, said yesterday in a phone interview from Zurich.

It’s simple, really. Policymakers continue to mess around with the rules frequently. Algorithms can’t keep track of the changes—they can’t judge, or intuit, the way humans can (the same emotional factor which, previously, allowed algorithms to dominate the markets).

For example, the Mexican peso used to be a measure of risk sentiment. It used to shift in line with the S&P 500 Index. Now it doesn’t. What the algorithms cannot do is account for the innumerable factors—tangible and intangible—that the human mind plays with (consciously and subconsciously) as it undertakes the decision-making process.

HFT’s time may finally be at an end. Reuters reports that Ewald Nowotny of the European Central Bank recently called for an outright ban on HFT practices, and The Bureau of Investigative Journalism reports that Andrew Haldane of Bank of America made similar comments last year.

The research is underway, but the attitude of skepticism toward HFT and the dominance of computers in the trading room is what’s really needed. At the end of the day, investors prefer a modicum of certainty rather than extreme volatility.


If you liked this article, you may also enjoy: