Hype over revolutionary new technologies such as AI is not a new phenomenon. Neither is the fact that they, counterintuitively, tend to make for lousy investments.
Take a look at insurance and global trade in the 1720s, canals in the late 1700s and early 1800s, and railroads in the mid 1800s. Or electricity in the late 1800s, the huge advances in intangible capital – measured by patents – in the 1920s, and the internet in the 1990s. All saw their fortunes follow a similar bubble-like pattern: An initial sharp rise in stock prices causes investors to pile in, driving down the cost of capital, innovative companies usually fall back to Earth, and many investors end up taking losses.
One of the hardest parts about watching this pattern persist is that inexperienced investors tend to play a key role; by chasing hype-related returns they push prices ever higher and ultimately end up on the losing side of the trade.
Novice investors tend to be more prone to biases that lead to chasing high returns (potentially contributing to the formation of bubbles), more prone to extrapolating recent price movements, more heavily influenced by attention and more prone to herding.
The basic idea is that investor excitement causes stock prices to rise above their reasonable fundamental values, often accompanied by a “new era” or “this time is different” narrative – something we have heard frequently about AI.
And when the hype for the latest technology trend dies down, asset prices are likely to follow. Suffice to say, I don’t think investing in AI companies is a sensible path to capturing the benefits of a potential AI revolution.
But AI is also drawing attention for another reason – the notion that it is getting so powerful that it could be used to beat the market. In other words, you don’t need to buy AI stocks, but maybe you can use AI to select a portfolio of stocks that will outperform.
Unfortunately, there’s a fundamental problem with this proposition. To win at investing you can’t just be good, you must be better than the competition. But as more skilled or AI-enabled competitors compete to find winning trades, any advantages conferred by AI decline.
That’s what appears to be happening with the AI Powered Equity ETF (AIEQ), which was launched in 2017. It “utilizes IBM Watson to equal a team of 1,000 research analysts, traders and quants working around the clock,” according to a description of the fund on etfmg.com.
Despite how impressive this sounds, the fund has not done anything special. It has trailed the market measured by both returns and risk-adjusted returns since inception. The reason is simple: A team of 1,000 research analysts offers no advantage over the collective knowledge of the market – especially when everyone has increasing access to AI tools.
The same conclusion can be drawn from a sample of 15 AI-powered mutual funds, including AIEQ, that were studied in a 2022 academic paper. The funds do not outperform the market on average.
All these funds have, or had, relatively high fees. Interestingly, only four of the 15 funds studied in that paper are still open today. Funds typically close after poor performance deters investors.
It’s not surprising to see funds like this closing. Long-term winners win not by outsmarting the market – whether using AI or not – but by making fewer mistakes than the competition.
The simplest ways to avoid mistakes are to minimize costs and own securities that broadly reflect the market. Chasing high returns from the next revolutionary technology is much more likely to be a mistake.
Benjamin Felix is a portfolio manager and head of research at PWL Capital. He co-hosts the Rational Reminder podcast and has a YouTube channel. He is a CFP® professional and a CFA® charterholder.