Are exchange glitches the price we pay for HFT?
The associated infrastructure demands of increased message traffic are a major factor in the more frequent exchange glitches that have occurred in recent years. High-frequency trading (HFT) relies on a greater flow of information to arbitrage across venues and exchanges have sought to tailor their exchange technology towards these participants to generate revenues. While the impact of HFT on institutional investors’ trading strategies is hotly debated, it is a growing segment of market activity that supplies the buy-side with liquidity.
The stresses placed upon exchanges to handle the substantial growth in message traffic is the cause of many of these recent outages. The need for exchanges to upgrade their platforms to better handle this growth in traffic also leads to teething problems that result in glitches.
The effect of HFT on execution quality, however, is an issue of greater importance to the buy-side, and has driven the growth in off-exchange trading for asset managers to effectively trade in blocks without HFT players getting in front of orders. This has in turn led exchanges to court HFT firms to fill a void created by dark pool trading.
Are compliance and stress testing standards sufficient to ensure participants and exchanges are protected in case of an infrastructure error?
With increased stress placed upon new and constantly updated trading platforms, exchanges struggle to iron out all kinks and prepare for all scenarios of trading. But market participants have shown a willingness to take matters into their own hands.
In the case of Nasdaq’s August failure, whereby trading was stopped for three hours, it appears market participants were adequately prepared for the disruption as none registered significant losses as a result of the Nasdaq trading halt.
The growth of exchange glitches feeds the shifting responsibilities of trading venues and buy- and sell-side firms. Pre-trade risk checks and automated checks to reduce ‘fat-finger’ errors have contributed in re-framing how participants deal with the possibility of infrastructure disruptions.
Does the risk of an exchange glitch put the onus on buy- and sell-side firms to ensure they have doubly strong internal systems to insulate themselves from market errors?
To a certain degree, market participants’ expectation that unplanned outages that affect their orders will occur. This shift in thinking has been driven through the electronification of trading processes and growth of highly-automated tools like algorithms that require complex risk mitigation and testing. Although not related to an exchange failure, the Knight Securities algo glitch of August 2012 that cost the firm US$400 million has sent an adequate warning to the market to ensure their systems are robust.
Nasdaq’s August glitch, in addition to its Facebook IPO issues last year, and other glitches on the Chicago Board Options Exchange in August suggest technology problems are likely to be with us for some time.