Isn't HFT Getting A Little ... Nuts?

HFT EvolutionWall Street & Technology's April 2013 digital issue takes an in-depth look at how high-frequency technology firms are using microwave and other technologies to survive in a challenging market. To read more, download our April 2013 digital issue now.

When high-frequency trading technology started to gather attention back in 2006, the standard measurement was in milliseconds. Quickly, the measurement for what was acceptable latency in the U.S. equities market dropped to microseconds by 2008, and by 2011 some extreme HFT shops were measuring response times in nanoseconds.

For a frame of reference, a microsecond is equal to 1,000 nanoseconds and 1 millisecond is equal to 1 million nanoseconds. So, to say that HFT is a million times faster today than it was in 2006 or 2007 would be fairly accurate.

In the HFT game, the number of big-time players seems to be thinning. It's hard to quantify how many shops are in the market, but we do know that the barriers to entry are becoming higher while the amount of money to be made from high-frequency trading falls.

For starters, share volume and volatility in the U.S. equities market continues to drop. These metrics play two very important parts in an HFT strategy: With higher volumes and greater volatility, there are more opportunities to profit. U.S. equities volume is down 33% since 2008, according to TABB Group. Although the research and advisory firm estimates volume will increase 6% this year, that doesn't come anywhere close to the all-time volume peaks in 2008, according to TABB's "U.S. Equities Market: 2013 State Of The Industry" report.

Despite the lower volumes, HFT activity will increase this year, likely approaching 52% of overall market volume, TABB says. While this is a slight increase from 51% of market volume in 2012, it's still short of the 61% that HFT controlled in 2009.

However, while HFT controls more than half of the overall market volume, industry revenue from high-frequency strategies will drop in 2013, down to $2 billion, according to TABB's report. This is a further decline from the estimated $2.2 billion in 2012 revenue, and a marked drop from the $7.2 billion in 2009.

To recap: Volumes are declining and revenue is harder to find, while it's getting more costly to compete in the HFT market. That doesn't sound like a good business to be in. In fact, in order to compete, HFT strategies must be the absolute fastest in the market. To achieve this, technology has to be constantly refreshed and bleeding-edge technologies are always being tested. Transmitting data between market centers via microwave data transmission is now in use between New York and Chicago, resulting in much faster speeds than optical lines but at a much higher cost.

[HFT: Makings of an Evil Empire]

At what point do the costs to compete become too much? For now it seems that the few large, ultralow-latency trading shops will continue to raise the technology bar. But at some point, maybe not this year or even next, it will become too costly to maintain the world's fastest trading technology, just to capture a piece of the shrinking HFT revenue pie.