Does HFT actually affect market quality at all? How do HFT activities influence market quality? It is significant to market regulators investors, and academic researchers to answer these questions considering the technological development and exponential increase in the sophistication of HFT strategies.
While interpreting the speed advantage of high-frequency traders merely as an informational advantage, HFT is just another form of informed trading which improves the price discovery process. This speed, however, may be used for other purposes as well. The decline in transaction costs due to extensive utilization of technologies is one of the valuable functions of HFT. While reducing errors and letting machines oversee everyday trading activities human traders can enhance attention spans in implementing their trading strategies.
Recent studies indicate that HFT can be helpful in improving market quality. However, the market impact of HFT during times of extremely high market volatility is unclear. After all, market failures and extreme volatilities existed even before the prevalence of HFT.
General Effects of High-Frequency Trading
Researchers have found that HFT activities decrease spreads, excluding a few exceptions. It is seen in certain markets that passive HFT activities lower effective spreads. HFT activities are negatively related to bid-ask spreads using international datasets from Tokyo and London.
Effective spreads increased significantly when trading activities of intensive algorithmic traders (iATs) decreased using data from the Toronto Stock Exchange. HFT has also resulted in narrower spreads which attribute to decline in trade sizes.
HFT activities do not always reduce spreads. While examining the technological upgrades on NASDAQ in 2010 which enabled faster transfer of messages, it was found they did not have significant effects on quoted and effective spreads. Analysis of the NASDAQ OMXS 30 index shows that competition among HFT firms induces more liquidity consuming trades, thereby draining available liquidity in the market.
On the contrary, Australian Securities & Investments Commission (ASIC, 2013) examines the Australian equity market and shows that HFT activities are not associated with changes in quoted depths at the inside spread, suggesting that liquidity is not deteriorated by HFT activities.
HFT activities also impact price efficiency as seen by researchers. Using the NASDAQ Data it is seen that price efficiency is positively associated with HFT aggressiveness. HFT trades tend to change in the direction of permanent change than in the direction of transitory movements while improving overall price efficiency. Price efficiency is positively associated with high-frequency quoting activities and algorithmic trading is associated with higher price efficiency.
In the three foreign markets: euro-dollar, dollar-yen, and euro-yen, this price efficiency has been measured by the frequency of triangular arbitrage opportunities and the auto-correlation of high-frequency returns.
An HFT proxy called as “strategic run” has been developed, which is often described as “series of submissions, cancellations, and executions that are linked by direction, size, and timing, and which are likely to arise from a single algorithm”.
Using this proxy it has been found that the short-term volatility of NASDAQ stocks, defined as the mid-quote range scaled by the mid-quote average during a 10-minute interval, declines as HFT activities increase.
This, however, does not imply that HFT activities can help prevent sudden market failures such as the Flash Crash. There also are findings where market-making HFT activities reduce short-term volatility (measured by one-minute midpoint quote changes) using the data of 30 stocks from the NASDAQ-OMX Stockholm.
An international analysis of 42 equity markets shows that short-term volatility (measured by standardized intraday price ranges) increases when the intensity of algorithmic trading rises. It should also be noted that the increase in volatility cannot be attributed to faster price discovery or to the penchant of algorithmic traders for entering volatile markets.
Some studies suggest that HFT firms are associated with order anticipation activities. Aggressive selling activities of high-frequency traders are generally followed by those of non-high-frequency traders, and the pattern remains same up to five minutes. This phenomenon exists due to order anticipation strategies of high-frequency traders. It is also suggested that high-frequency traders employ order anticipation strategies in the E-mini S&P 500 futures market.
High-Frequency Trading and Market Glitches
Under normal market conditions, HFT activities are not harmful to market liquidity. Several market glitches have been attributed to HFT, although not many academic studies find HFT to be detrimental to market quality.
Nearly one trillion U.S. dollars’ worth of equity value vanished within a matter of minutes during the Flash Crash of May 6, 2010, with the Dow Jones Industrial Average temporarily plunging by more than 9%. The market did not take much long to recover from the crash; the event has raised concerns about the market stability and cause of such sudden market failure.
The negative media coverage of HFT and the Flash Crash raised significant interest and concerns about the fairness of markets and the role of high-frequency traders in the stability and price efficiency of markets. To study this, the U.S. Commodity Future Trading Commission and the U.S. Securities and Exchange Commission (CFTC and SEC, 2010) analysed two datasets, the FINRA and the Lit Venue Datasets.
FINRA Dataset investigation suggests six of the 12 high-frequency traders have reduced their involvement in the market sometime after the crash. This caused a decline in overall market liquidity. Staffs of CFTC and SEC examine the Lit Venue Dataset during the crash find that high-frequency traders were engaged in aggressive selling activity. These trading activities fell while recovery which followed the crash. HFT was not responsible for causing the crash as concluded by Kirilenko.
Is the current market condition more susceptible to events like the Flash Crash or recent market glitches considering the prevalence of HFT? This is a very natural question. It would be only fair to say that possibility of such events existed before the rise of HFT as well. Some discontinuities have always occurred in markets even before the age of electronic trading, while dislocations are harmful to market integrity; as flickering quotes have existed well before the advent of high-frequency quotation.
Liquidity providers can always exit without notice, exposing marketable orders to price risk, if liquidity provision is not mandated by law.
There can be two illustrations given of the events similar to the Flash Crash:
- Kennedy slide of 1962 – On May 28, 1962, due to an unknown cause stock markets in the United States experienced a sudden turbulence and some stocks fell by more than 9% within 12 minutes. If you had never noticed it, that is because it occurred not in 2010, but in 1962.There was no HFT during this age! The SEC hoped to find out the causes of any wrongdoing in order to pass legislation to prevent further malfeasance but did not find any such evidence of such behaviour.
- Human Errors before HFT: In the 1990s a clerk was fired for mistakenly trading two million U.S. dollars’ worth of a stock, instead of the intended two million shares. This clearly suggests that trade mistakes among skilled traders existed even before the advent of HFT.
Empirical evidence to date generally suggests that high-frequency trading has improved market quality during normal times. What is not clear is the role of high-frequency traders during episodic periods of the market crash and extreme volatility.
If you’re a retail trader or a tech professional looking to start your own automated trading desk, start high-frequency trading training today! Begin with basic concepts like automated trading architecture, market microstructure, strategy backtesting system and order management system. You can also enroll in EPAT, one of the most extensive algo trading courses available in the industry.
Disclaimer: All data and information provided in this article are for informational purposes only. QuantInsti® makes no representations as to accuracy, completeness, currentness, suitability, or validity of any information in this article and will not be liable for any errors, omissions, or delays in this information or any losses, injuries, or damages arising from its display or use. All information is provided on an as-is basis.