Standard Life Aberdeen harnesses artificial intelligence to pick stocks
When human beings falter, try a machine. That seems to be the approach at Standard Life Aberdeen, which is turning to artificial intelligence to run a new investment fund that it hopes will grow into a $10 billion blockbuster.
The investment manager has announced plans for the use of machine-learning to identify which shares to buy. There will be no human input or an override function for its decisions.
Standard Life Aberdeen has teamed up with a technology institute run by Mitsubishi UFJ, the Japanese banking group, to develop the technology for the new fund, called the Aberdeen Global Artificial Intelligence Global Equity. Standard Life Aberdeen and Mitsubishi are each putting $5 million of seed capital into the project, which executives hope could become one of the biggest funds in its stable.
“We would hope that this is a $10 billion fund,” David Wickham, global head of quantitative solutions, said. The fund will be marketed to institutional and retail investors.
Other fund managers have tried to harness the investor appeal of AI by creating funds focused either on so-called Faang stocks — that of Facebook, Apple, Amazon, Netflix and the Google owner Alphabet — or on the crunching of “big data”, the ability to access and manipulate huge amounts of figures, for new money-making insights. Standard Life Aberdeen, however, believes that there is more potential in applying AI to its factor-based investing division, which runs $49 billion of client money.
Goldman Sachs and Blackrock have launched funds aimed at applying AI to big data, but Mr Wickham said that this approach was overhyped. “Our view is that big data won’t lead to big alpha.” Alpha is the extent to which individual stocks outperform share market indices.
Factor-based investing, sometimes known as “smart beta”, is based on the philosophy that over the long term certain types of stock persistently outpace others. Value stocks beat growth stocks, small companies beat big ones and companies that have done well in the recent past beat those that have not.
However, these factors fade in and out. The machine’s job is to identify patterns of the past to predict when those factors are likely to gain traction in the future, and to time investments accordingly.
Tests of the system suggested that it would outperform its benchmark, the MSCI All Country World Index, by a “punchy” 4 per cent to 6 per cent a year, Mr Wickham claimed.
Andy Haldane, chief economist at the Bank of England, predicted this week that AI and the automation of cognitive skills would have a greater impact on the economy than the first industrial revolution, when steam power transformed the country.
Full article: The Times, 22 August 2018