The buzz around artificial intelligence and the role it could play in revolutionising the asset management industry grows ever louder. But not everyone is convinced.
Despite the long queues at recent fintech trade shows in Hong Kong and Singapore to highlight the tech's potential, not everyone is rushing to jump on the AI bandwagon, concerned that humans are not yet close to a position where they can truly harness its power.
“Where there are errors occurring in decisions made by AI technology, humans still struggle to identify the reasoning and logic behind it,” Lake Chen, quantitative portfolio manager at CSOP Asset Management in Hong Kong, told AsianInvestor.
It is partly for that reason that there are still relatively few AI-supported or AI-managed funds in Asia. CSOP, certainly, is is no hurry to launch AI-managed funds, Chen said.
However, lured by the potential rewards that come from having first-mover advantage and the cutting-edge experience that sometimes only comes from being able to learn on the job, a few pioneering souls are pressing ahead.
In March this year, the Hong Kong-based arm of mainland fund firm China AMC created a multi-factor managed account operated via AI.
ChinaAMC teamed up with Hong Kong-based fintech firm Magnum Research, whose Aqumon automated-investment engine was created at the Hong Kong University with funding from the Alibaba Entrepreneurs' Fund.
ALLURE OF AI
Other specialist funds are emerging too, including Finex Asia, a Hong Kong-based manager of AI-driven funds investing in US consumer loans.
In August this year, Finex Asia started offering Hong Kong investors an AI-driven fixed income fund product that pays a monthly dividend. The firm buys consumer loans from the US market, partnering with online platforms such as Lending Club, Prosper and Upgrade and uses AI to enhance the yield by picking out the best loans for the fund portfolio. The AI is a refinement of the proprietary model built by Dianrong, a mainland Chinese technology platform.
The structure of such AI-driven products potentially points the way ahead for the asset management industry and shows how new technology could enable fresh ideas to be offered to investors at a lower cost.
The first generation of AI-managed fund portfolios show promise but have been found wanting in times of high market volatility.
In response, some local regulators have already begun to formulate new rules to deal with a possible new generation of tech-enabled investments.
The Hong Kong Securities and Futures Commission (SFC), for example, issued new guidelines for crypto funds this month. In the announcement, the SFC's chief executive Ashley Alder also addressed the issue of AI-run products and the need to look at their governance structures and risk management.
“The type of fintech which promotes financial inclusion and efficiency needs to be distinguished from emerging technologies whose value to society is yet to be proven. This is particularly so when these technologies raise potential financial stability concerns and pose immediate investor and consumer protection risks,” he said.
AI is slowly making inroads though. Among the apparent ways AI can help with portfolio management, for example, is to refine quantitative models or strategies using exchange-traded funds (ETFs), providing investors with cheaper and more effective exposures.
China AMC's head of ETFs, Frederick Chu, told AsianInvestor that the aim of their AI-driven managed account is to be able to identify the most effective factors to harness excess returns. Using a machine-learning stock-picking algorithm, it applies a smart beta-like multi-factor approach, making changes according to different market cycles.
“It's only a matter of time before AI massively disrupts the nature of the products being created for a wider range of investors,” Finex Asia co-founder Maggie Ng told AsianInvestor.
What's more up for debate is how much time it will take.
Finex’s Ng acknowledges that the market for AI-driven funds is still new in Asia and will take time to develop. “Asian investors are not aware of this consumer loans platform, nor this type of asset,” she said.
On the other hand, later entrants could find it harder to compete: "The cost of entry is high; to build an AI model is not an overnight thing. You can build it in a couple of months, but to train the AI model takes years."
In the meantime, as AsianInvestor has reported, machine learning is being used to utilise data for improved marketing and client relationship management.