(Pictured: Zak May)
by Penny Pryor
High frequency trading (HFT), and its analysis, continues to court controversy in the US and globally.
One of the most comprehensive global studies done on potential winners and losers of the practice, first released in 2010, has recently been republished with the express authorisation of the Commodity Futures Trading Commission (CFTC) in the US.
The Risk and Return in High Frequency Trading paper by Matthew Baron of Princeton University, Jonathan Brogaard, from the Foster School of Business, University of Washington and Andrei Kirilenko from the MIT Sloan School of Management, was previously authorised for public disseminating by the CTFC Office of General Counsel and circulated as “The Trading Profits of High Frequency Traders.”
This research was suspended when the CTFC’s outside academic research program itself was suspended in 2012, following a complaint by the world’s largest derivatives exchange operator CME Group. CME Group alleged that researchers were illegally accessing sensitive market data for high-frequency trading research. An internal investigation at CTFC found that this was not the case.
In it’s original form the paper found that high frequency traders make, on average, US25c per trade.
“This equates to $18,799 per day for each HFT in the August 2010 E-mini S&P 500 contract alone,” the paper said.
In its second incarnation the paper focused more on the winner take all attitude of HFTs and the relative profits of the types of high frequency traders.
The latest release of the paper expresses the cost to the non-HFT trader in basis points.
“Finally, the decomposition of profits also shows that despite strong outperformance on the individual firm level, effective HFT trading costs imposed on non-HFT investors are only 0.22 basis points,” the authors found.
Kirilenko says the changes to the latest edition were made to make the paper tighter.
“Research economists look for specific things in specific format in specific places of the paper and if they don’t find it there or find something else instead, they get impatient or suspect.
“Probably our colleagues wanted to see percentages rather than dollar amounts so the results could be compared to other assets or markets – just a guess. The paper will certainly continue to evolve in response to comments from our peers,” he said.
Kirilenko, who is a former chief economist at the CFTC, also recently testified at a US Senate Committee hearing on HFT.
In Australia the Australian Securities and Investments Commission (ASIC) collects audit trail data on Australian trading activity, which could be used for similar analysis of the practice here. However this analysis has not been undertaken or has been undertaken and not released, despite requests from industry associations like Industry Super Australia.
“We asked them to do an analysis like this when they did their taskforce,” Zachary May, director of policy at Industry Super Australia, said. “This analysis could go to the heart of investors’ concerns about the fairness of the current market structure. It is powerful stuff.”
In a market supervision update in April ASIC said: “Currently, for example, we are looking at a high-frequency trading participant who appears to be regularly putting in ‘bait and switch’ type orders. Using Market Analysis Intelligence, we have identified hundreds of such instances by this participant’s facilitation desk.”
That suggests it does analyse the data to some extent.
“A responsible analysis of audit trail data that explores whether retail investors systematically lose out to HFT could put some real numbers around the costs,” May says.
The table below shows the costs per trade from the US study. It analyzes the decomposition of average daily short-term profits among different trader types in August 2010.
Small traders, with a median trading volume of less than 20 contracts, lost a total of $2.6 million during the month. But the $31 million lost by opportunistic traders is also significant as that category captures medium-sized buy-and-hold investors.
The labels HFTA, HFTM, and HFTP correspond to Aggressive, Mixed, and Passive HFTs, respectively.
Source: Risk and Return in High Frequency Trading