The title of this article is borrowed, in part, from the third instalment of the Terminator movie franchise, released in 2003. The movie tells the story of a robot that returns to the year 2018 from a post-apocalyptic world with intent to destroy humanity’s ability to stand in the way of a future ruled by hostile machines by taking control of world’s computer systems. Far fetched? Perhaps, but there are some interesting similarities to that storyline with what researchers and market watchers are starting to observe looking at the gyrations taking place across the world’s financial markets over the past few years.
In May 2010, all the key US market indices experienced a massive collapse and then a swing to recovery, close to previous levels, in little more than half an hour. All this happened in the early hours of the morning before the trading floor was even open. The Dow Jones in particular experienced a loss of 1000 points in just a few minutes, equating to trillions of dollars in value. In 2013, a similar crash was experienced on the Singapore exchange that lost nearly US$7 billion in market capitalisation over 3 days. More recently in August 2015, another flash crash took place in the US with another 1000+ point drop precipitated by sell-offs in China and index drops in Europe prior to the opening on trade in the US. Major stock markets are not the only ones affected. At the start of 2016, Bloomberg reported on the flash crash of a 10% drop in the value of the South African Rand in the matter of minutes of one morning.
There is a lot of speculation as to the causes behind these changes from highly publicised examples of individual rogue traders all the way through to the panic-induced retail investor. As debates have evolved and research has discovered, there is an increasingly common denominator starting to emerge. A CNBC report from January 2016 recognises the US trading regulator CFTC’s (Commodity Futures Trading Commission) findings that “One of the key connections is the rapid placement and withdrawal of trading orders that lies at the core of high frequency trading (HTF)…this activity was at least significantly responsible for order imbalances in the derivatives market, which in turn affected the stock market.”
The Rise of the Machines
So where are the evil robots? The same CNBC report goes onto point out: “These crashes are caused by institutional trading from exchange traded funds (ETFs) and HFT. They are not caused by mums and dads trading because mums and dads simply do not act in such a coordinated fashion in such a short timeframe. Mums and dads also do not have the leverage to shift markets in this way within 30 minutes or an hour. That power lies in the hands of large-scale derivative traders.”
Large scale derivative trading is an increasingly automated process. Traders, ultimately human decision-makers, develop and instruct automated systems to make decisions on their behalf in times, places and frequencies that would not be humanly possible. Powerful institutions equipped with the capital, skills, information and infrastructure seek and extract return across global capital markets. As has been identified, the losers in this war are the retail investor and those institutions that are unable to compete with the sophistication and speed that these powerful robots have at their disposal.
It is not the robots that are evil, it’s the set of rules and algorithms that determine their behaviour to buy or sell that are the real problem. These rules do not apply discretion beyond what they are coded to do. Scarily, they are able learn by these rules and market movements to serve their masters for effectively and efficiently in future. Unsurprisingly, it’s the masters of the machines that stand accountable.
A relational view on robotic trading
The world’s financial markets are a clear example how relationships between nations, regions, industries and asset classes can impact each other through complex interconnections. Robotic trading is a mechanical attempt to identify these connections and derive an information advantage in a market to deliver financial return by trading assets based on rule-based assumptions. Essentially, this mechanical process looks to transform the observed dynamics of a relationship into a tradable transaction for gain for those who command it.
The unsettling reality is that investigators and researchers still do not know to what extent robotrading affects the world’s markets because of the pervasiveness of systems and the herding effect that they induce when it comes to other institutional investors trying to follow the lead of their more sophisticated peers. What is clear though, is that there are casualties in the process. Mums and dads do not have the speed and information to make decisions like robots do, but there still remain many millions of families across the world that invest their savings into markets to pay for school fees, retirement and rainy days. Attempting to make sense of market information so skewed by machines can be a fearful and potentially fruitless pursuit.
Colin Habberton is CEO of PayProp Capital in Stellenbosch in South Africa. Before that he was CEO of the GivenGain Foundation South Africa (GGFSA).