Forget HFT; “High Intelligence Trading” is the new frontier for technology, markets, regulation
By Henry Engler, Compliance Complete
NEW YORK, Apr. 10, 2014 (Thomson Reuters Accelus) – While fast is good, smart is better, and with untold resources of computing power and memory banks in the clouds, the new frontier in electronic trading combines sophisticated intelligent software with rapid-fire processing, enabling traders to stay one step ahead of the regulators.
“What’s the difference between pure speed and adding intelligence to that speed?” asked Terry Keene, head of iSys, a technology integration firm, at a conference focused on high performance computing. The answer is “big data analytics” that brings decision-making and trading to a “near-time” environment, he added.
Companies small and large, including IBM, are at the forefront of leveraging so-called “big data” and ever faster and more powerful computing power. The outcome for market participants, particularly buy-side firms, is an environment that makes millisecond trading decisions more dependent on analytics and information scoured from a variety of sources.
“Over the past 18 months people started to understand data,” said George Kledaras, chief executive of Fix Flyer, a technology company with a focus on the FIX protocol, a long-standing communications standard developed for securities transactions and markets.
“If I can predict that I can be the first in the (trading) queue 80 percent of the time, that’s good enough,” he added. “I don’t know if it’s illegal, or whether the government will make it illegal.”
Power Linux: IBM’S super-computing platform
When IBM’s “Watson,” the world’s most advanced question answering machine, beat a pair of seasoned “Jeopardy” players on the popular game show in 2011, the technology backbone to the artificial intelligence was fueled by the company’s Power Systems servers. These high-octane servers are now being used along with Linux, an open source operating system that allows independent developers to contribute to its growth and evolution. Together, the PowerLinux servers enable the handling of large quantities of data more efficiently without the need for additional hardware.
Many smaller technology companies are partnering with IBM to leverage PowerLinux for trading applications that bring together often diverse pools of data. In financial markets, the application can be used for many purposes, such as scouring Twitter feeds for information on trends in the real estate markets. More controversially, perhaps, this type of technology can also be used to handle news feeds containing sensitive government economic and policy news, and parse natural language within milliseconds to effect trading decisions.
FOMC “No taper” decision
Last September, when the Federal Reserve’s Open Market Committee announced that it was not going to “taper,” or reduce its bond-buying purchasing program, hundreds of millions worth of stock traded within a millisecond before the embargoed 2 p.m. release time for the Fed’s announcement. What was most striking was the fact that these trades were executed both in Chicago and New York at the same time, raising speculation that the Fed’s decision was known beforehand by high-frequency trading firms.
While the Securities Exchange Commission and Fed have launched investigations into the trading activity – the results of which have yet to be announced — the event also put a light on the release of market-moving information through news agencies and their proprietary feeds, as well as the technology employed by trading firms to parse natural language in milliseconds. Industry experts say that even if the Fed’s policy decision was not leaked or known beforehand, a simple headline that read “NO TAPER” could have been easily interpreted by such software, unleashing massive pre-programmed execution orders and leading to spikes in bond and equity prices.
From a regulatory perspective, participants at the conference believed that attempts to limit trading firms from using such tools to their advantage was likely to be a losing battle, with advances in software outpacing the ability to understand how they were applied.
“You are not going to stop them,” said Keene of iSys. “It’s not going to stop. It’s an ecosystem that keeps evolving.”
(This article was produced by the Compliance Complete service of Thomson Reuters Accelus. Compliance Complete provides a single source for regulatory news, analysis, rules and developments, with global coverage of more than 400 regulators and exchanges. Follow Accelus compliance news on Twitter: @GRC_Accelus)