Robo-advisors have traditionally been perceived as asset allocators, providing a pool of ETFs, or in some cases mutual funds, to construct risk-based portfolios for clients. Generally, robos are perceived as disruptors of mainly retail distribution, with asset managers directly providing their products to the robo platforms.
However, some robo-advisors are now able to provide stock-picking strategies for clients.
For example, Magnum Research, the firm behind B2C robo-advisory platform Aqumon, and China Asset Management in Hong Kong two years ago launched an AI-powered China A-share portfolio.
The product, the AI China Alpha Strategy, is only available to professional investors and it now has $100m in assets, according to Don Huang, the firm’s co-founder and head of quantitative research.
Magnum provides the software that constructs the portfolio, while China AMC is the named portfolio manager that executes the trades based on the instructions it is given, said Huang.
With stock-picking and asset allocation capability, Magnum will have the capability to debut its own AI-driven funds. To that end, the firm received a Type 9 (asset management) licence from the Securities and Futures Commission in September.
Ken Shih, the firm’s head of sales and marketing, told FSA recently that his firm does plan to launch its own funds at some point. But operating as an independent asset manager, additional investment and operational capabilities would be needed.
“Our bread and butter right now still remains on how we can work with others to provide value, instead of doing it ourselves,” he said.
Shih explained that at the moment, the firm’s focus is providing its robo technology to partners, which include asset managers and banks. He also noted that Aqumon couldn’t have launched the AI China Strategy alone, as China AMC provided the investment knowledge of China A-shares when constructing the strategy.
Other robo-advisory firms in the region have started to launch stand-alone strategies, though they only use ETFs. Singapore-based Stashaway, for example, launched in September an income portfolio that invests equity, bond and Reit ETFs.
The AI China fund
Out of the 3,000 stocks in the China A-share universe, the Magnum/China AMC AI China Alpha Strategy only picks around 50-100 names, according to Huang.
The strategy operates like a multi-factor or smart beta fund, in which it looks at 250 factors, including financial information, market sentiment and macro-economic data. It also applies machine learning, which means the technology decides by itself which factors and stocks should outperform in a given market environment.
“Typically, different factors will contain some clues for a stock’s future return. But it will be very hard for human beings to detect the relationship between all of these factors in a huge investment universe,” Huang said.
He claims that Magnum’s technology is able to assess the relationships between the different factors from one another and how these relationships change over time as the market environment changes.
“It is a dynamic model, in which machine learning can learn different patterns and [attempt to] predict future returns.”
The model also takes into consideration portfolio risks, such as individual stock concentration and sector risks.
In terms of performance, Huang declined to provide specific figures, citing regulatory requirements about disclosing private fund information.
However, Huang claims that the strategy has outperformed the market. Since its inception in January 2018, the strategy has been up by at least 30%, versus the CSI 300 Index return of -6.68%.
In 2018, although the strategy was down, it had at least 20% outperformance versus the China index, according to Huang.
