We learnt more about artificial intelligence (AI) and its use in building a model portfolio - or even a practical portfolio for folks to invest in - on April 3.
First off, for old-fashioned investors who have followed the teachings of investment guru Benjamin Graham - and this includes Warren Buffett - there is no need to wonder if value investing is still relevant. It is. And AI will add to value investing.
Man + Machine
According to Paul Ho, the ever-youthful looking head of investment technology at UOB Asset Management (UOBAM), who could easily pass off as a young man in his 30s, Asian investment managers would have to sift through some 25,000 stocks in Asia to find the best stocks.
Most investment managers have a team of five.”We would all cover 50 stocks each for a maximum of 250 stocks. How many stocks are there in Asia including China, Japan, and India? 25,000. That's 1% coverage. Most of the time most stocks go uncovered. Which one 1% do you choose to cover?” Ho wonders.
After meeting with Larry Cao, formerly a senior director of industry research, CFA Institute, who now focuses on AI research and its integration into finance and investments, UOBAM started to incorporate aspects of AI into its investment strategies and this enabled the UOBAM team to actually sift through these 25,000 stocks.
See also: Pictet filing mistakenly showed US$15 bil bet on Chinese bank
“We constructed a model that looks at the 25,000 stocks, as well as the macro data such as GDP, money supply, technical analysis etc. We evaluate all these stocks and rank them by expected return in the next one month,” Ho explains.
See also: S'pore's net AUM inflows falls to $193 bil in 2023, but AUM grows to $5.4 tril
“In our model, we process more than 30,000 variables at the same time which is significantly more than the human mind can process,” Ho says. Permutations and combinations of 30,000 variables run into more than 100 million. None of the data points or variables are weighted. More than that, the algorithm used by the UOBAM is not proprietary but on open source and widely available.
Although other fund houses are also using AI, UOBAM is likely to be doing theirs differently as a couple of its unit trusts, United Asia Fund and United Greater China Fund have outperformed the benchmark MSCI Asia ex-Japan Index and peers in the one year, three years and five year categories.
“The algorithm we’re using has been around for a long time, 20 years and has been used successfully in many industries including by names such as Walmart, Amazon and Netflix,” Ho says, adding that the portfolio is dynamic rather than static.” We have about a 20% turnover a month,” he indicates, referring to the number of stocks that are changed.
