- AI can pick private equity firms to back better than many investors, per an Oxford study.
- Researchers trained AI to read 400 prospectuses to predict the top-performing funds.
- However, lead researcher Ludovic Phalippou believes it should still only be used as a sounding board.
Artificial intelligence can pick out top-performing private equity (PE) funds better than many institutional investors, according to research from the University of Oxford.
Researchers trained an AI model to identify high-performing PE funds by reading 400 fundraising prospectuses and factoring in long sections that are often overlooked by investors, for a study by the Saïd Business School.
The system selected firms that outperform peers by around 5% each year from a pool of European funds. Using a standard metric that considered both released and unrealized returns, the top quartile of funds selected by the AI model generated 2.10x the original investment versus an industry average of 1.85x.
During the due diligence process investors, or limited partners (LPs), typically look at PE firm reputation and past performance but this is ultimately unrelated to fund performance, the study found.
It described qualitative information in fund prospectuses, specifically the long strategy section where fund managers describe the investment opportunity they are targeting, as salient but overlooked. One reason for this could be that qualitative is not readily comparable from one fund to the other, and not easily digestible, the study said.
Ludovic Phalippou, the lead researcher and a professor of financial economics at Oxford Saïd, said that LPs shouldn’t use AI as a final decision-making tool but as a sense-check.
“You cannot automatize it and not have any humans intervening,” he told Insider. “That’s a recipe for catastrophe.”
If it tells an LP ready to cut a check that there’s a 0% chance of the chosen fund outperforming, they should reconsider, and vice versa, Phalippou said.
Still, investors must factor in the wider macroeconomic environment before making decisions. In a buzzy market, LPs are in high demand. AI could also help streamline the work of sifting through pitches, he said.
“It is the first time that artificial intelligence is shown to be able to select investments based on dense, and technical documents that are written exclusively for highly specialized and large institutional investors; and the robot beat most of these large professional investors,” Phalippou said.
It follows a report from PE firm Coller Capital, a secondary buyer that has $27.5 billion in assets under management, which found LPs see AI uses in each stage of the PE transaction process. Respondents found AI “significant” in sourcing deals, assessing them, and working with portfolio funds once deals are done.
AI in investing is not new; its use for intense data analysis in stock trading is well established. It allows a huge amount of data to be crunched and modeled, to ultimately help investors come up with a strategy.
But only a limited number of investors have the resources to take advantage of AI, so Phalippou expects to see consultants add it to their services.
Read the full article here