Skip to content
  • Friday, December 20, 2024
  • Home
  • FCA says HFT firms exacerbate and profit from flash crashes

FCA says HFT firms exacerbate and profit from flash crashes

  • July 6, 2018

A paper published by the UK’s Financial Conduct Authority has said that high-frequency trading (HFT) firms and agency traders exacerbate flash crashes in the equity market, with HFT firms profiting from the events.

The research note entitled “How do participants behave during Flash events? Evidence from the UK equity market,” was written by Matteo Aquilina, Jia Shao and Carla Ysusi from the Economics Department and Brian Eyles from the Secondary Market Oversight Department of the FCA.

It looked at mini flash-crashes/flash rallies, which it defined as large price movements that revert within a short time window and during we see high levels of traded volume occur. The study looked at extreme events for the 18 months between January 2014 and June 2015 in FTSE 350 stocks. These mini-flash crashes/rallies occur more frequently than large 2010 Flash Crash, 2014 US Treasuries flash crash or the 2016 Sterling flash crash but are smaller in scale.

The researchers broker market participants into proprietary HFT firms, hybrid firms which “mainly provide agency trading services but may use a similar technology to HFTs or provide DEA to HFTs” although largely not via proprietary trading and non-HFTs, which is a proxy of trades done on behalf of institutional investors.

The FCA looked at 33 HFTs, 29 hybrid firms and 190 pure non-HFTs. It noted that HFTs account for approximately 30% of the traded volume, hybrids for approximately 60% and pure non-HFTs for less than 10%, based on previous research.

It found that hybrid firms appear to be driving the extreme price changes by trading most aggressively and building large positions, while HFTs add some pressure by trading aggressively during the price spike. Overall their magnitude for direct trading is considerably smaller than hybrids. Pure non-HFTs are not “main players” during these events.

During the recovery period after the event, HFTs submit more liquidity than the amounts that are consumed by other market participants. Hybrids do not resubmit enough liquidity to replenish the amounts that are being consumed by other traders, even if in most cases they submit higher volumes than HFTs.

However HFT liquidity is not submitted at the top-of-the-book or close to the best bid/offer. For shocks lasting longer than a second, HFTs resubmit liquidity far from the prevailing price. In some cases they even move some of their resting liquidity from the top levels of the book to lower levels.

“All these behaviours are likely to exacerbate the price movement as they affect the prices at which participants can trade,” notes the report.

In shocks lasting longer than a second, hybrids tend to resubmit liquidity at the top or mid-levels of the book, however, they do not resubmit enough quantities to substitute the liquidity taken from them.

“Overall, the depth of the order book is much thinner by the time the price starts to revert and even after the event has ended, because of the behaviour of both HFTs and hybrids,” noted the authors.

The authors added that they are unable to distinguish the types of order flow underlying hybrid transactions which could be driven by their own low-latency proprietary desks, agency business and the type of orders offered to their clients or other firms trading through them using DEA.

“Given what we know about HFTs through the result for the HFT category, it is less likely the main results are driven by HFTs trading through them using DEA,” wrote the authors.

While the report calculated that hybrid firms make a loss in these events, it found HFTs on average make small profits.

“This is probably due to their superior ability in speed which allows them to better manage their risks and avoid adverse selection,” it concluded.

No conclusive triggers for the 2010, 2014 or 2016 flash crashes have been established by authorities, despite investigation.

©TheDESK 2018
TOP OF PAGE