The Gateway to Algorithmic and Automated Trading

The future of HFT? More risk and fewer easy pickings

First Published 12th February 2013

It hasn't taken long for the HFT industry to get nostalgic about the good old days. Adam Cox reports on the views from one panel discussion.

London - The easy pickings don't look so easy these days.

High frequency trading firms may still be making money, but it's getting tougher all the time and that's a trend that is unlikely to reverse, industry watchers say.

"The low hanging fruit has been plucked," said Donald MacKenzie, a professor of sociology at the University of Edinburgh who has been studying the HFT industry.

MacKenzie, speaking during a panel session at this week's High Frequency Trading World conference in London, cited estimates from Rosenblatt Securities which concluded that high frequency profits probably fell to about $1.25 billion in 2012. That was down from nearly $5 billion at a peak in 2009.

"Whenever an industry has grown quite quickly, people just presume that it always will, and that's not a reasonable assumption," said fellow panel member Patrick Boyle, a founding partner at Palomar Fund Management.

For a short period of time, people certainly were cleaning up, Boyle added. "Essentially they took a trade which used to be done manually, they automated the decision making and the speed of execution and, you know, the manual guys were horrified - they couldn't get a fill, they didn't know what was happening."

But that inevitably had to change.

Boyle offered some historical perspective. Initially, he said, financial firms would create high-frequency trading teams with, say, 30 people. But perhaps half of them would take signing bonuses and go off and start teams elsewhere.

"Suddenly every firm on the street has a team of guys that's promised them risk-free return, and they're all running the identical strategy and that can't possibly work," he said.

At the same time, firms were throwing money at technology and did not necessarily allocate costs correctly. So they might have a large team that was making 20 million a year, but not factor in that they were spending 15 million of that on technology.

"You have, you know, hundreds of small firms that were doing quite innovative things but I think almost that that speed of growth is what wiped out a lot of the profitability in the sector. A lot of the trades just aren't that scalable. You look at most HFT firms - the capital they need is for capex, not really for trading capital," Boyle said.

An example of how aggressively the financial industry piled onto the HFT bandwagon could be seen in the way that exchanges courted high frequency firms.

Boyle said a decade ago there may have been about seven or so different order types, whereas now there could be 500.

"These are all order types that are there for HFT traders," he said.

"Being an expert in market microstructure and having these types of orders that the market didn't know about obviously drove a lot of profitability a few years ago. But the problem is that whenever something's extremely profitable, and also extremely low risk, it essentially can't survive. Any arbitrage has to eventually return to the risk-free rate."

During the session, the panel was asked whether regulatory uncertainty could be temporarily holding back HFT firms. MacKenzie said his impression was that that was not the case.

Both MacKenzie and Boyle predicted more consolidation among HFT firms.

Professor Dominic Swords, a research group director at Henley Business School, moderated the discussion. In an earlier presentation, he offered his view about the risks the HFT world faced.

He likened the current environment to what had happened in the mountain-climbing arena. Whereas before it might take three days to scale Eiger, now the climber Ueli Steck can do it in two hours and 47 minutes.

"He broke the mould and transformed the way in which we attack that," Swords said.

But to do that, he had to strip away the gear he used to the absolute minimum. "The same metaphor transfers to high frequency trading. Because you are a slim, agile, efficient operator, there are certain risks," Swords said.

One of those is periodic illiquidity, with a second risk coming from the imponderable nature of what happens when algorithms interact with other algorithms. "You can get a small change in the market that can become amplified into a major discontinuity by the way in which algorithms leap on top of each other and magnify the change," Swords said.

Finally, there was the question of behaviour.

"Clearly, in a system where profits go to those players with greater insight, greater access to better information and speed of response ... that will give people an ability to manipulate the market inappropriately," he said.