Algorithmic stock trading rapidly replacing humans
Algorithmic trading, including high frequency trading (HFT), is rapidly replacing human decision making, according to a government panel which warned that the right regulations need to be introduced to protect stock markets.
Around one third of share trading in the UK is conducted by computers fulfilling commands based on complex algorithms, said the Foresight panel in a working paper published yesterday.
Nevertheless, this proportion is significantly lower than in the US, where three-quarters of equity dealing is computer generated.
The Foresight panel, led by Dame Clara Furse, the former chief executive of the London Stock Exchange, argued that there are both benefits and severe risks to algorithmic trading.
There was “no direct evidence” that the computer trading in itself increased volatility, it said, but in specific circumstances it was possible for a series of events with “undesired interactions and outcomes” to occur and cause massive damage.
One such event is self-reinforcing feedback loops, whereby small changes, perhaps driven by data delays or news events, loop back on themselves and trigger a bigger change, which again loops back. Another event, normalisation of deviance, is more psychological: unexpected and risky events come to be seen as ever more normal until a stock market crash occurs.
Algorithmic traders’ regular dependency on limited capital while using ultra low networking latency systems presented risks to liquidity, and with the better predictability of order flows there were also risks of market manipulation.
On the positive side of electronic trading, the panel said, liquidity has improved, transaction costs are lower, and market efficiency is generally better.
The paper also stated that technological advances in extracting news will generate more demand for high frequency trading, but higher levels of participation will limit profitability.