A newly published research paper has cast doubt on the effectiveness of machine learning (ML) and artificial intelligence (AI) in equity trading by suggesting that humans are still more effective than their algorithmic counterparts when it comes to beating the stock market.
In particular, the study found that while a lot of academic research claims that ML and AI produces highly accurate forecasts and profitable investment strategies, these claims are not matched by “real-world AI-driven investments” which the study states is “ambiguous and lacking in high-profile success cases”.
The study, conducted by Barbara Sahaklan, Fabio Cuzzolin and Wojteck Buczynski and published in the International Journal of Data Science and Analytics, looked at 27 peer-reviewed studies on AI market forecasting conducted between 2000 and 2018.
It found that most of the studies cherry-picked data by running multiple versions of their investment models in parallel and presenting only the highest performing model.
“This approach would not work in real-world investment management, where any given strategy can be executed only once, and its result is unambiguous profit or loss - there is no undoing of results,” states the report.
The study also highlighted the lack of explainability in the majority of “black box” algorithms used for trading – an issue that has generated increased regulatory attention – as well as a lack of disclosure over performance. Furthermore, the small number of AI-powered funds that have disclosed their performance have tended to underperform the market.
These factors are cited as two reasons why AI-based trading has failed to attract widescale adoption.
“As such, we concluded that there is currently a very strong case in favour of human analysts and managers. Despite all their imperfections, empirical evidence strongly suggests humans are currently ahead of AI.”
The study’s authors do recognise some advantages from AI technology that could be applied to the investment world. Firstly, it states that there is a “promising case for using ML in combination with alternative data” to predict movements in local markets.
Secondly, it states that AI’s ability to process huge amounts of data can “add tremendous value in the investment process” and in ESG screening given the disparity among ESG data vendors and providers.