What does an asset manager do when one of its actively managed equity mutual funds falls flat with investors? In the case of J.P. Morgan Asset Management, it shakes up things by infusing artificial intelligence and machine learning into the investment process in an effort to construct a winning stock portfolio that attracts investor interest.
The company today publicly unveiled the JPMorgan U.S. Applied Data Science Value Fund (JPIVX), a product formerly known as the JPMorgan Intrepid Value Fund. This large-cap value fund began trading in 2003, and according to Morningstar its performance has slightly exceeded its large-value category peers during the three-, five-, 10- and 15-year periods. But it has also underperformed its benchmark during these time frames. (Morningstar uses its own US Large-Mid Cap Broad Value Total Return Index as the benchmark, while JPMAM uses the Russell 1000 Value Index. Either way, the fund has underperformed both indexes over long-term periods.)
While the fund has $243 million in assets, it seemed that JPMAM viewed it as a laggard. Thus, the company felt it was time for a different approach. On July 1, the fund dropped its Intrepid Value Fund name, became the U.S. Applied Data Science Value Fund and began to employ a strategy that couples traditional active management with insights using data science.
“We wanted to bring a differentiated alpha and portfolio construction approach to our clients in the Large Cap Value space and the fund was not growing so we selected this strategy to convert,” a company spokesperson said via email. She didn’t comment on why the company waited so long before announcing the change.
In a press release announcing the new-look U.S. Applied Data Science Value Fund, JPMAM’s head of U.S. structured equity, Hamilton Reiner, noted the company in recent years has bolstered its traditional fundamental investment process with newfangled data science capabilities such as AI and machine learning.
“The investment process is driven by machine learning and works off the core belief that there is significant alpha potential in portfolio construction, creating value through both security selection and allocation decisions,” he said.
The press release also stated these efforts have culminated in the creation of a new business unit focused on the application of AI and machine learning.
One of the U.S. Applied Data Science Value Fund’s three portfolio managers, Eric Moreau, moved over from the data science team to help run the fund, the spokesperson said. Moreau and the other two portfolio managers, Wonseok Choi and Andrew Stern, work with the company’s data scientists as part of the investment process.
According to the prospectus, the data science approach uses proprietary techniques to process, analyze and combine a wide variety of data sources. That includes J.P. Morgan’s proprietary fundamental research, company fundamentals and alternative data. The latter can include unpublished data unique to a company’s prospects such as global supply chain data, news feeds and social media.
The combined input is used to identify companies that fund managers believe have attractive valuations relative to their perceived risk levels.
The fund had 91 holdings as of the end of November, with top sectors being financials, healthcare, industrials and information technology with weightings that ranged from 21.5% to 10.2%. The top five holding were NextEra Energy Inc., Bank of America Corp., Prologis Inc., Alphabet Inc. Class C and BlackRock Inc.
Though it’s a small sample size, as of yesterday the fund had gained 5.7% since it changed its format and name on July 1, compared to a gain of 3.5% on the Russell 1000 Value Index.
Meanwhile, fund fees were trimmed by 10 basis points on the Class A, C and I shares, according to the company spokesperson.
The growth of big data analytics has permeated spheres ranging from sports to the financial services industry, and AI and machine learning have become more prevalent among asset managers looking for an edge in a very competitive space. It’s something that JPMAM plans to expand upon going forward.
“We would expect to launch additional [AI and machine learning] capabilities in the future, but [we have] no details to provide just yet,” the JPMAM spokesperson said.