Can AI Replace Your Fund Manager? In Time, Yes.
By Lewis Braham
Aug. 23, 2023 2:00 am ET
Artificial intelligence replacing human beings has been the subject of science-fiction books for decades. The distinctly human part that the AI usually can’t replace is often the “emotional” one, i.e., the soul.
That’s bad news for fund managers who often pride themselves on unemotionally pursuing profit and analyzing financial data objectively, without any all-too-human cognitive biases. They seem prime targets to be replaced by soulless AIs that not only objectively analyze millions of data points per second but also execute an effective investment strategy based on them.
So far, there are only 11 AI-run exchange-traded funds, based on a Barron’s screening. Others, though, are probably employing AI behind the scenes. “If I were in an asset manager’s shoes right now, I’d be all over looking at AI, but in more of the perspective of how do I strip some cost out of my business and make it run more efficiently?” says Lee Davidson, Morningstar’s chief analytics officer.
The good news—sort of—for human managers is that index funds are still the primary threat, as the current AI-run funds have also struggled to beat their benchmarks. Only one, Qraft AI-Enhanced US Large Cap QRFT –0.19% (ticker: QRFT), stands out for performance. But that should change in time, as more funds launch and the AIs “learn.”
Analyzing stock data with computers is hardly new. It’s called quantitative investing. But AI is about teaching the computers to think for themselves and adapt to changing market, business, or economic conditions. That could ultimately replace even the qualitative analysis of a company’s business that human money managers still do.
AI is already being used to monitor earnings conference calls with analysts, says CEO Ryan Pannell of Kaiju Worldwide, which subadvises an AI-run ETF called BTD Capital DIP –0.39% (DIP). The AI will be “listening in real time,” he says. Not only will it analyze the text of the executives’ speech for key performance-influencing words, but also “monitor the tenor, pitch, and pace of the CEO’s voice. From that, it will be able to determine the likely market reaction to the strength of the [CEO’s] comment.”
Today’s AI-run ETFs aren’t nearly as advanced. BTD Capital has a focused “buy on the dips” strategy for stocks. The challenge, Pannell says, is “to identify authentic dips—a price retracement, which is artificial in nature. So, if you buy it, there will be a mean reversion from low to high, and you’ll benefit from that little pop in price action to the upside.”
The problem is, some stocks can have industry- or company-specific declines and not recover. Such was the case in March, when Silicon Valley Bank went bust and other California banks dropped. “The regional banking collapse was a pattern that the AI had never seen before,” Pannell says. That lack of understanding hurt the ETF, and it is underperforming this year. But the ETF’s AI has now “learned what that pattern looks like,” Pannell says, so it can mitigate losses in the future.
Qraft AI-Enhanced US Large Cap employs an adaptive AI system to adjust its exposure to five major stock factors—value, quality, size, momentum, and volatility—depending on market conditions. It has beaten 70% of its peers in Morningstar’s Large Growth category in the past three years, without human managers.
Francis Oh, CEO of Qraft Technologies, which runs four factor-based AI ETFs, says convincing mom-and-pop investors to trust a machine is difficult; the funds only have $30 million in them. “Humans still have a tendency to believe the decision making of humans, regardless of whether they are making good returns or not,” he says.
It also took decades for investors to accept the automated index fund, invented in 1971. Given Wall Street’s love of AI, acceptance will probably come sooner here.