Harvard study shows AI stock trading rivals many picks made by fund managers

Some bad news for all the mutual fund managers out there: A new study from researchers at Harvard Business School seems to support the fear that artificial intelligence and machine learning could do their jobs. But here’s the catch—with only about 71% accuracy, depending on how predictable their trades are.

The working paper “Mimicking Finance” from Lauren Cohen, Yiwen Lu, and Quoc H. Nguyen, published this month by the National Bureau of Economic Research, finds “that 71% of mutual fund managers’ trade directions can be predicted in the absence of the agent making a single trade.”

The paper goes on to say, “For some managers, this increases to nearly all of their trades in a given quarter. Further, we find that manager behavior is more predictable and replicable for managers who have a longer history of trading and are in less competitive categories.”

What does that mean? Basically, that the trades of more senior managers, especially those who are in less competitive areas, are easier to mimic (and thereby, those jobs might be easier to replace with AI).

The findings are based on data the researchers analyzed from 1990 to 2023 that took into account the size of the fund, the broader economic indicators, and investor flows.

Perhaps what’s most alarming for mutual fund managers, though, is the paper’s conclusion: “For some managers,” AI predicted “nearly all of their trades in a given quarter”—which is the equivalent of a mic drop.

However, there are a few big caveats. The paper finds that the larger the ownership stake of the manager in the fund, the less predictable their behavior. It also found less predictable managers strongly outperform their peers, while the most predictable managers significantly underperform.

Even within each manager’s portfolio, the research shows “those stock positions that are more difficult to predict strongly outperform those that are easier to predict” (a bright spot for fund managers who want to keep their jobs).

The study is significant because it explores which tasks could be automated using AI, and how that could affect jobs in the financial sector. It estimates the U.S. asset management industry to be worth about $54 trillion.