AI shifts sector’s future
• Can technology boost investment returns to sustain longer retirements? asks Pedro van Gaalen
Numerous technology-mediated factors are converging to reshape the broader investment industry. As an example, ongoing medical advances continue to increase longevity, which means retirement investments will need to stretch further to meet an investor’s future cost of living. Fortunately, developments in the application of technology in asset management and investing are helping to keep step with these changing demands.
Artificial intelligence (AI), for instance, will profoundly impact how risk is priced and investment returns are achieved, says Trevor Abromowitz, Boutique Head of Liability-driven Investments at Old Mutual Investment Group.
“When we live longer, our liability increases — we effectively need more capital now to fund our cost of living in the future.”
The question is, can AI boost investment returns to sustain longer retirements? One local fintech start-up believes so.
Launched three years ago, Stellenbosch-based NMRQL Research has developed an AI-managed fund that uses machine learning (ML) algorithms and big data to make predictions about financial markets and inform investment decisions.
Stuart Reid, chief scientist at NMRQL, explains that technology removes the inherent human biases and emotions that exist in traditional investment strategies, which assumes that future prices are driven by certain risk factors.
“As such, asset managers try to tilt their portfolio to exploit these factors. The assumption is that this investment approach will be rewarded with a return in the future. However, it is prone to the fallibility of human cognition, which can be influenced by factors such as confirmation biases, anchoring and the fact that humans are constantly fooled by randomness. Also, there is no feedback loop in this process.”
Conversely, NMRQL’s ML system leverages built-in feedback mechanisms, applying algorithms — NMRQL currently has 1,200 in production — to more than 19,000 big data sets, which can include diverse data points such as financial reports, news, social media and sentiment data, to continually adjust its “beliefs”.
“As the model receives feedback, it retests its ‘assumptions’ and learns something new, which it then applies to investment decisions going forward,” explains Reid.
This happens on a weekly basis as NMRQL doesn’t target high frequency trading. “Our strategy is to apply technology to make informed long-term predictions, as we believe longer frequency investments deliver greater returns.”
This approach also mitigates the impact of transaction costs which, when combined with the economies of scale these systems can achieve, will lower the cost of investing.
“This is another significant contributor to delivering better returns over investment life spans, as technology is already helping to bring down the hard pass-through costs associated with regulatory compliance and fund administration.”
Abromowitz adds that the applications of AI-enabled technology in other areas of the sector are also reshaping the investment value chain, which can create additional operational and cost efficiencies.
“Chatbots, for example, are being used to engage with customers to meet certain service requirements, while robo-advisors can now provide financial advice with little or no need for human intervention.”
However, Abromowitz doesn’t foresee a scenario where AI disintermediates the human aspect completely.
“Savvy asset managers who embrace AI are likely to deliver better outcomes than those who continue to apply either human intelligence or AI in isolation. However, the ultimate success of this combined approach depends on the process employed,” he suggests.
Reid agrees some asset management models will seek to augment and improve the human element with AI.
“While I believe that ML will replace unsophisticated human investors, it will also be a boon for intelligent investors who can leverage the technology to amplify their alpha, reduce their costs, or test their assumptions about the world. That is why we’re also catering to the demand for augmented decision making through our decision-support business. This offering provides access to our bespoke, customised platform to help investors make more informed decisions. Rather than retrospectively rate risk, the data analytics provided will help to shape opinion and improve predictions by adding datadriven insights to the human decision-making process.”
But, he adds, not every fund that applies AI will use it successfully, nor is there any guarantee that the data provided when augmenting AI into the process will influence the decisions asset managers take.
While AI can more accurately rate risk and generate better returns for investors, Abromowitz believes the true benefit of technology can only be realised when investing behaviour is altered.
“While these applications are innovative and effective at delivering returns, AI is unlikely to meaningfully boost savings allocations. It is ultimately up to investors to find more efficient ways to build up capital needed to invest and benefit from these technological advances. Only then can they expect to sustain their retirement and meet the increased liability associated with a rise in life expectancy.”
ASSET MANAGERS WHO EMBRACE AI ARE LIKELY TO DELIVER BETTER OUTCOMES THAN THOSE WHO APPLY AI IN ISOLATION