As the number of domestically listed exchange-traded funds (ETFs) surpasses 1,100, the focus of investment is shifting from 'what to buy' to 'how to select.' With the rapid increase in products, investment decisions have become more complex, prompting the introduction of a 'three-step strategy' that utilizes artificial intelligence (AI) to help investors select, compare, and invest in suitable ETFs.
Choi Chang-kyu, head of ETF Research at Mirae Asset Management, presented this strategy during a keynote speech at the 'ABC Launch Celebration AI Ecosystem Innovation Forum' held at The Plaza Hotel in Jung-gu, Seoul, on July 9. He emphasized that AI can be a key tool not only for selecting and comparing ETFs but also for validating investment strategies.
The first step is 'selecting.' This involves filtering ETFs that have a high allocation to desired investment themes, such as AI or semiconductors. He stated, "With over 1,000 listed ETFs, the first step is to narrow down the options. If you want to find a semiconductor ETF with a high allocation to SK Hynix or SK Square, you can download the Portfolio Deposit File (PDF) data and ask AI to analyze the top 10 holdings' weights, sector weights, and recent changes in holdings to efficiently identify the desired ETFs."
The second step is 'comparing.' By inputting the ETF's PDF into AI, investors can easily grasp the composition of holdings, sector weights, and any overlapping or differentiated components among ETFs. Choi noted, "Many ETFs have similar names, and there are also U.S.-listed ETFs, so comparison is always necessary. Creating tables to compare top holdings, sector weights, and overlapping stocks makes it much easier to determine which ETF aligns with your investment style."
The final step is 'investing.' AI can simulate various investment scenarios, including buying points, asset allocation, rebalancing, and loss tolerance. Choi himself uses AI to analyze the diversification of ETF portfolios, maximum drawdown (MDD), and Sharpe ratios. He mentioned, "During the week, I run out of tokens, so I come in on weekends. I use the leftover tokens to create ETF portfolios and simulate diversification, MDD, and Sharpe ratios. This simulation is one of the areas where AI is best utilized in the ETF industry."
Looking ahead, he identified promising investment themes such as AI infrastructure, humanoid robots, and aerospace and defense. However, he clarified that AI should not replace investment decisions. Choi remarked, "When you ask ChatGPT to recommend a good ETF, it may suggest ETFs that are not listed domestically or even products that are not from our company but claim to be. AI should be used as a tool for data analysis, not as a decision-maker."
He added, "AI responds based on what I know. It is much more effective to use it for analyzing top holdings, sector weights, and overlapping stocks rather than asking it to predict which ETF will rise. Ultimately, it is up to the user to effectively utilize AI, and developing the ability to leverage AI is crucial."
* This article has been translated by AI.
Copyright ⓒ Aju Press All rights reserved.