Equities trading technology has evolved rapidly in the last ten years or so. To such an extent that many are asking if a “tipping point” has now been reached. A tipping point where it is the machines that now dominate the markets rather than humans.
This might seem like science fiction but it is in fact a very valid question. Computer algorithms now account for anything up to 80% of total US equity volume, according to the High Frequency Trading Review.
So what are the dangers associated with this paradigm? Could these computer algorithms go wrong and start misfiring? Could they flood the markets with rogue orders that trigger a catastrophic meltdown?
The answer is yes, possibly. But how real is the risk?
In order to understand the risks, it is important to understand how these high frequency trading algorithms work. In essence, they take information in from a variety of sources (news feeds and price feeds for example), perform various computations based on continuously maintained parameters and spit out orders into the electronic markets, often thousands of order per second.
The algorithms themselves follow all kinds of pre-defined strategies. Some strategies look for extremely short-term arbitrage opportunities whereas others are more directional in nature. Often, the algorithms will continually scan and sweep the various trading and execution venues looking for the best price in a particular stock. These days, there are many such venues where liquidity in a particular stock might be available, including not just the exchanges but also a number of so-called “dark pools”, i.e. crossing networks operated by various independent entities where liquidity is not displayed.
The risk around these trading activities comes from the fact that the algorithms can generate orders at such high frequency and that those orders are sent into the markets automatically. So what if one of these algorithms were to go wrong? If someone mistakenly sets a wrong parameter on one of these systems, the potential is for the algorithm to continuously send erroneous orders and trigger a chain reaction that sets off all the other algorithms in an unstoppable flood.
The regulators are still trying to get their heads around all of this, which is why the SEC recently published a consultation paper, seeking input and feedback on a range of proposals to better control high frequency trading.