Asset managers are making big commitments to research in and development of artificial intelligence and data science techniques, spending they believe is necessary to remain viable, executives said on a panel Monday.


Speaking about lessons from the great recovery at the Milken Institute's 2018 Global Conference, Emmanuel Roman, CEO of PIMCO, said, "for big long-only businesses, part of the solution is data and the ability to acquire new ways to look at the issue of [alpha]." PIMCO, for instance, needs to acquire and make sense of every new data set that becomes available on the manager's huge mortgage business to accurately predict trends such as pre-payments, which are key to determining portfolio value. "We need a lot of data and we have to find the right data," Roman added. This won't be cheap. PIMCO is hiring 250 people in technology, opening an office in Austin, TX, and making other capital investments to support the effort.

[II Deep Dive: The Michael Milken Project]


It's not only asset managers for public markets — which have long been defined by data — that believe AI and data are key to creating excess returns.

Jonathan Lavine, co-managing partner at private equity and credit firm Bain Capital, said, "the businesses we invest in will be impacted by technology." Internally, the firm has collected 30 years of data on its own investments and now needs to better understand that information to analyze where it's been successful and where it could do better, Lavine added. "But understanding that [from a data perspective] doesn't substitute for judgement... That's particularly true where we loan money to a lot of small businesses that have a story. AI won't get that story for a long time." Lavine stressed that it will take longer for the value of data in private investments to become clear. But, he said, "people who are resistant and aren't willing to experiment will be left behind." 

GoldenTree Asset Management founding partner and chief investment officer Steven Tananbaum agreed. "You wouldn't want to compete against somebody who is fluent in data but you're not," he said. For example, as an investor in Puerto Rico, GoldenTree needed to determine the number of residents who have left the island. Public reports ranged from 100,000 to 500,000. Yet GoldenTree found that 80,000 had actually relocated using cell phone information, according to Tananbaum. "That is an example of data that wouldn't have been available a decade ago," he said. 

One longtime industry veteran harkened back to Wall Street's last technological revolution, in the 1980s. At that time, Barings CEO Thomas Finke pointed out, the financial industry made a lot of money trading government bonds. That market has become increasingly electronic, wiping out jobs and profits in the process. "In the short term, data is very disruptive to asset management, but long term it should benefit investors and issuers," Finke said.


But AI and data are just tools, the money management executives said. They don't change their underlying business. "It doesn't replace bottom-up analysis," PIMCO's Roman said. "It's a tool, but a very useful one."

II's Julie Segal will be reporting from the Milken Institute's Global Conference throughout the week.