Since its inception in 2015, robo-investing has emerged as a tool with the potential to minimize the impact of market volatility in South Korea.
The profitability of this investment method -- based on computerized data analysis, not human instinct -- has been questioned. But the time has come, unexpectedly, to prove who is correct. The catalyst is the coronavirus outbreak that has been rattling the stock market at home and abroad.
According to market data tracker FnGuide, 17 robo-investing funds in Korea took an average of a 2.44 percent loss from January until Monday. In the past three months, with COVID-19 on the wane, the robo-investing funds, which mostly invest in foreign products, gained 8.2 percent.
The combined volume of robo-investing funds -- created by asset management firms on the basis of automated trading algorithms -- came to 96.6 billion won ($79.83 million) as of Monday in the nation, where investors poured in some 270 trillion won in public funds. The volume shrank by 6.8 billion won in three months as retail investors in robo-investing funds cashed out their investments during the recovery.Human intervention at play
While the figures indicate a solid performance by robo-investing funds, it appears that the robo-investing mechanism still hinges on the human element.
Automated as the funds are in terms of investment decision-making processes, it is human asset managers who construct portfolios and control factors, and investors still have to decide when to buy and when to exit from the funds.
This, along with other factors such as target asset class, has resulted in a discrepancy in yields between automated funds.
Nine out of 12 Korean AI funds -- excluding those that either are younger than 6 months old or have accumulated less than 1 billion won -- took losses from January until Monday, with the greatest loss amounting to 8 percent, the FnGuide data showed. Half retreated over 2 percent, while four out of 12 underperformed the average of active funds -- 2.85 percent loss -- investing in a mixed bag of foreign stocks and bonds.
In the meantime, algorithms themselves appear to have stopped short of whetting investor appetite for such funds. None of the 10 feeder funds, which had pooled retail investors’ money publicly for at least a year, were managing less than 5 billion won.
If a fund more than a year old is smaller than 5 billion won and fails to attract additional investments for at least a month, a fund manager may opt to liquidate the fund under the Korean rules. This has led some of the AI-powered fund managers to scrap their products later this year, including one created by NH-Amundi Asset Management and advised by robo-adviser engine developer December & Company.
Experts say artificial intelligence itself is not a fortuneteller that can predict which stocks or bonds will perform better, and is not designed to fulfill that function.
“(Robo-investing funds) seem to have failed to differentiate themselves from ordinary active funds or quantitative funds,” said Kim Woo-chang, associate professor of industrial and systems engineering at the Korea Advanced Institute of Science and Technology. “Speaking of profitability, robo-investing funds have yet to prove that its mechanism is superior to other traditional methodologies to manage a fund.”Beauty of AI-powered financial advice
The question is where robo-advisers’ deep learning mechanism could come into play and have an edge over other funds.
Some robo-adviser engine developers here argue that there is still room, saying the beauty of AI lies in its ability to predict when to buy or sell assets, instead of where.
To implement AI-powered strategies, robo-advisers can bring in the algorithm that focuses on when to buy or sell, inspired by the concept known as “gamma” first introduced in 2013 by US financial services firm Morningstar.
In this regard, there needs to be a clear distinction between creating a robo-investing fund -- focusing on automated choice of assets (“alpha”) and asset allocation (“beta”) -- and acting as a robo-adviser to offer a tailored portfolio depending on individual risk tolerance and other factors -- focusing on advising on the timing.
“Decent financial advice from robo-advisers can make better portfolio decisions, which could lead to potential benefits to an investor’s return on assets,” said Julius Chun, head of robo-adviser engine developer startup Doomoolmori.
Chun added that AI’s role is crucial in that sense, because such investment decisions are inferred and reached from myriad cases and experiences -- an area where big data analysis jumps in.
Robo-advisers’ offers of investment advice powered by AI is in contrast with robo-investing funds, given the required human efforts from both buyers and sellers in funds.
“Robo-investing funds are financial products to sell, whereas robo-adviser services are not,” December & Company CEO Chung In-young said. “Once the financial product is sold, individual investors who bought the product are responsible for taking care of the aftermath, because financial institutions’ role is limited to attracting retail investor money.”
Chung added that robo-adviser services should strive to offer fully personalized investment advice to each individual, because otherwise the service will be no different from robo-investing fund products.More access to investors
For the past couple of years, Korea has moved to allow companies with AI technology to manage financial consumers’ money via online, as long as the 1.5 billion won capital requirement is met, through a series of deregulations.
Still, from robo-advisers’ perspective, there is a long road to go to achieve independence from financial institutions and offer full-fledged financial services to address each individual’s needs.
One could be to join a deregulation initiative by the government, under which a robo-adviser startup is expected to hone its methodology to analyze an individual’s characteristics as an investor.
Chung of December said the company is looking to participate in a state-led pilot program called the “MyData” initiative, through which a licensed MyData operator is able to request the financial data of third parties.
“An individual’s characteristics should be measured based on the data analyses, in addition to surveys,” Chung said.
Chun of Doomoolmori said Korean robo-advisers are subject to tougher regulations, compared with those in the United States and the United Kingdom.
“As a robo-adviser, there are numerous constraints as to how to reach out to a wider range of individual investors,” he said.
Chun added that robo-advisers’ targets of investment are limited to exchange-traded funds, equity-linked securities and public equity funds. They have limitations when directly investing in equities.
Kim of KAIST pointed to the power dynamic between large financial institutions with sales networks and robo-adviser service providers without such networks.
“Without direct channels to reach out to a number of financial consumers, robo-advisers will remain underdogs,” Kim said. “But I believe the government is open to deregulations to encourage more financial technology firms to enter into an untapped realm.”
By Son Ji-hyoung (email@example.com