Fidelity Investments is turning to its internal massive data and behavioral science to drive a competitive edge, as the asset management giant lays the groundwork for systematic funds.
The firm’s head of quantitative research and investments Neil Constable told MarketWatch in an interview that he hired Gilbert Haddad, who previously worked at Steven Cohen’s Point72, to lead Fidelity’s investment decision science team. Haddad started at Fidelity last week, according to Constable, formerly head of global equity at Jeremy Grantham’s GMO.
The effort in decision science is new at Fidelity and draws on the activities of its fundamental portfolio managers and analysts, according to Constable. He said the firm, which oversees $4.2 trillion of discretionary assets, has enormous internal data on trading activity, analyst ratings on stocks and bonds, the events surrounding investment decisions — and the outcomes.
“Think of it like playing ‘Moneyball’,” said Constable, referring to the title of Michael Lewis’s book on the important insights helping to drive success in baseball. “It’s really applying behavioral finance and behavioral science to understand” what the portfolio managers and analysts are good at or not.
The goal is to “help them understand their own biases” so they can strengthen their investment processes, Constable explained. He also aims to use the collective information arising from Fidelity’s fund managers and analysts to create “signals” used in new “quantitative products.”
Fidelity’s scale and decades of internal trading and analysis data give it a “unique” informational edge in the world of asset management, where fund managers can access the same external alternative data, according to Constable. He expects the team led by Haddad will initially consist of a few associates and grow from there.
A spokesperson for Point72 declined to comment on Haddad’s departure.
Fidelity’s Constable spent several years studying math and science before starting a career in finance.
After receiving an undergraduate degree in physics and math from the University of Calgary in 1996, he went on to earn his master’s degree in applied mathematics from the University of Cambridge in 1997. He then earned a PhD in physics from McGill University in 2001, and later that same year became a postdoctoral fellow at the Massachusetts Institute of Technology.
Constable entered finance in 2004 as a quant researcher at State Street Corp. He then worked almost 13 years at Boston-based GMO before becoming chief investment officer of San Francisco-based CircleUp in 2019. He joined Fidelity in July 2020 to lead its quantitative research and investments group, or QRI.
“This job did not previously exist at Fidelity,” said Constable, who reports to Bart Grenier, the head of the firm’s asset management business. He said Grenier created QRI last year to bring together fixed-income and equity quants, as well as data scientists who work in artificial intelligence and machine learning.
As the head of QRI, Constable oversees more than 80 people. The quant fixed income, quant equity, advanced data and analytics, quant index solutions and quant research services teams were combined under him to provide differentiated data and analysis to Fidelity’s traditional active management teams in equity and fixed income.
QRI’s second mandate, he said, is to develop systematic products driven by unique data sources for Fidelity’s customers.
“One of the reasons I was hired was to build out a systematic, scalable, ultimately very customizable product lineup for Fidelity’s client base,” said Constable. “To kick start that, we just announced a month ago that we’re going to be acquiring the active quant team from Geode Capital Management.”
Under the agreement, Fidelity is bringing over actively managed quantitative equity funds and Geode’s separately managed account business for a combined $15 billion of assets, according to Constable. Both businesses will report into QRI, he said.
Fidelity’s push to amp up its quantitative efforts comes as 2021 marks 75 years since the asset manager was founded by Edward C. Johnson 2d in 1946, according to its website. The firm’s scale across asset classes over decades, along with its technology, creates “opportunity all over the place,” in the eyes of Constable.
“The amount of internally generated data is huge,” he said. “So systematically collecting, organizing, analyzing and making use of that across all our investment products is a big competitive edge, especially as a quant.”
‘Data, data everywhere’
Data is a lot easier to come by today than when Fidelity began decades ago.
Nicholas Colas, co-founder of DataTrek Research, captured this in an Aug. 4 note in which he described a visit to Fidelity as a Wall Street analyst who began covering the auto industry in 1991. “Fidelity was the U.S. brokerage industry’s most important client,” he wrote. “That meant every Wall Street analyst was in and out of their offices at least once a quarter visiting analysts and portfolio managers.”
He recalled entering Fidelity’s “chart room,” whose walls were covered with “graphical representations of data” tracking everything from stocks to gross domestic product. Colas saw a “gaggle” of portfolio managers around one wall with graphs showing the weighting of each sector in the S&P 500 as well as other equity indexes going back several decades, according to his note.
“In pre-Internet times, this was not easy information to get,” he wrote.
“In the modern world, it’s data, data everywhere,” said Constable. “A lot of technology is needed to harness that data and turn it into actual information.”
Fund managers have been turning to so-called alternative data in a bid for market-beating gains. The data may be collected from online sources such as social media or come from “really obscure” or specialized sets from areas like the shipping industry and consumer retail, said Constable.
Fidelity’s advanced data and analytics team acquires alt data, much of which requires some form of AI, such as natural language processing, to organize and make it useful for investment teams, he said.
As for the “chart room,” Constable said it still exists at Fidelity, with another group maintaining it. In his view, the room, an “early manifestation” of the firm’s emphasis on data collection and analysis in active management, is “very consistent with the overall philosophy inside Fidelity.”
It also signifies how times have changed.
“In 1950, the cutting edge was collecting all that data manually and systematically putting it on charts so the investment teams could consume it for stock picking or bond picking,” said Constable. Now it involves computers collecting “every piece of random data generated by every industry or every government entity,” he said, and systematically “cleaning” it for portfolio managers to use.
Check out: How investors are using alternative data to track the inflation debate
But it’s Fidelity’s own data that will differentiate it from the quant competition as it pushes into new areas, according to Constable.
The asset manager already has some examples of funds that are “mostly systematic,” such as the Fidelity High Yield Factor ETF, FDHY,
QRI is also harnessing Fidelity’s internal data with the longer-term vision of running purely systematic funds in equity and fixed income, which eventually will become available to Fidelity’s retail customers.
“Purely systematic is on the horizon,” said Constable. “But we certainly established a beachhead in that direction.”
