The current era of investing, at least when it comes to the big boys such as asset managers, is one defined by algorithms, big data and automation. However, institutional investors have also realised that algorithms and big data approaches to taking investment decisions still have some weaknesses.
The human mind still has qualities difficult to replicate through AI. And perhaps the biggest issue with algorithms is that if every big investor is using them, they all cancel each other out and nobody benefits. The answer many are arriving at has been dubbed ‘quantamental investing’ and is an approach which involves both man and machine, each complimenting the other’s strengths and weaknesses.
JPMorgan Asset Management’s approach to quantamental investing is a new unit that has been named ‘intelligent digital solutions’. One of its projects uses an AI neural network to scan a huge database of historical earnings call transcripts. It hopes the AI will be able to pinpoint particular words used in those transcripts that represent a pattern of provoking positive or negative market reactions to an earnings call. Whenever a new earnings call includes words, or patterns of their frequency on combination that the neural network has highlighted as heralding a particularly positive or negative market reaction, its team of investment analysts will receive an alert.
Other quantamental systems in the works include one that Essentia Analytics is working on. The idea is that it will spot when fund managers are deviating from their usual investment system. Human psychology can influence even experienced professional investors and lead them into mistakes such as overtrading when things are going badly or hanging onto assets for too long in a bear market. The system will send fund managers an email alert when it spots them taking an approach to investing that doesn’t fit their usual pattern.
Other tech being worked on is a dashboard that classifies investment teams by their particular strengths and weaknesses such as, for example, good timing but less impressive portfolio construction. The dashboard can also highlight things like a strong stock picking record but sector biases. This knowledge will help institutional investors achieve more efficient division of labour in allocating different investment responsibilities between teams.
The quantamental approach to investment splits opinion. Technology advocates believe that algorithms taking decisions on huge historical data sets will inevitably perform better than humans and that all blending the two will achieve is a dilution of the edge given by strong algorithms. Other believe that tech-led investing is a fad and the dynamic and evolving nature of financial markets means human insight and intuition will always have its role.
Which camp is proven right in the long run, only time will tell!