Fans of Terminator will remember how machines dominated the earth, replacing humans in the series. With the rapid development in machine learning and big data techniques, one question that needs to be answered is “Will machine replace investment managers?” Alex Birkin, EY Partner, Global Advisory Leader - Wealth and Asset Management does not think so. While high frequency, short-term trading is almost fully automated, machines are not yet capable of taking correct long –term investment decisions independently, shared Alex.
Speaking at the seventh edition of Cafemutual Confluence, Alex talked about the recent developments in artificial intelligence (AI) and shared why a fully automated investment process is still many years away.
Why is it so? According to Alex, these techniques tend to over-fit data, create spurious patterns and correlations. In addition, these algorithms struggle in times of economic turmoil or during changing macroeconomic conditions and they are unable to capture complex human responses. Furthermore, it is difficult to integrate these techniques with the traditional investment management processes.
Despite this, acceptance of AI is increasing among asset managers. When Alex polled top hedge fund managers on their plans to include inputs from AI in their investment decisions, most of them responded in affirmative. In fact, a third of them had already integrated an element of machine learning in their investment process.
At the heart of it, machine learning helps asset managers spot patterns that humans can’t easily spot given the sheer amount of data being created in world today, said Alex quoting CEO of Man Investments. But what does it mean for alpha generation? Data shows that over the one, three and five year periods, hedge funds using some form of AI inputs have outperformed their peers using traditional investment methods.
What this means is that there is huge scope to use big data and AI to augment the decision-making process of fund managers by analysing vast quantities of data to identify patterns. Typically, the first movers will benefit the most. However, the end point that is full automation of the entire life-cycle of investment decision making is still many years away.