Elliot Noma has spent 30 years working in finance and nearly 50 years studying the mind while satiating his curiosities. At 65, he doesn’t appear ready to wind down. In fact, the next frontier to conquer will be in creating an AI-driven strategy for his fund. By Anthony Malakian with photos by Timothy Fadek
When you look up at the sun, can you describe how bright it is? Can you tell the intensities of brightness between today, yesterday and a week ago? Can you rank those differences? Do you think the differences in those rankings—and how others ranked that same brightness—says something about your personality? Do you think those results reflect what kind of trader you’d be?
These are the kinds of questions Elliot Noma has been asking of others since the early 1970s. When he was a student over four decades ago, he didn’t envision himself becoming a hedge fund manager, and he didn’t think that the work he was doing back then would still have such a tight correlation to what he does today. But, indeed, the path was being paved.
While these are questions he’s been asking for decades, it wasn’t because he envisioned himself as a hedge fund manager; it was because he was simply curious and a professor stoked those flames of interest. It’s a curiosity he’s looked to satiate his entire life. He’s been a student and a professor, a banker and a hedge fund manager, an entrepreneur and a tinkerer.
While Noma has largely spent his career in finance, this is only because the profession has allowed him to indulge his interests. He dabbles in advanced mathematics, psychology, sociology and technology. It ended up being a perfect concoction for a life in the capital markets.
The only question is this: When will this decades-long odyssey finally wind down?
“The idea is to figure out some way to build in greater automation so that a small staff can turn over these strategies and have a better, more automated system to develop strategies, screen them, and implement them quickly and responsibly. There’s a lot to be explored with AI and a lot that remains to be seen as to which areas of finance present the greatest opportunities.”
Elliot Noma, 65, is not the typical subject of a Waters magazine cover story. Usually the standard fare is the CIO, CTO or CEO of a large asset manager, hedge fund, bank or exchange.
What makes Noma different—and interesting—is his background and the fact that while he’s been inside the world of finance since the 80s, he’s always been on the move. You get the feeling talking to him that his main concern is finding a new challenge rather than working toward a specific title.
He currently runs a tiny hedge fund, Garrett Asset Management, which he started in 2008 along with Aliona Manvae. Garrett engages in systematic trading of futures, exchange-traded funds (ETFs) and currencies. At its peak, it managed less than $5 million, a relative pittance compared to previous Waters buy-side cover subjects like JPMorgan Asset Management, Vanguard, Blue Mountain, or Cerberus. The vast majority of this magazine’s cover stories profile the 1 percent of the financial services industry’s elite; Noma represents an everyman who has turned his curiosity into a unique entity.
School as Foundation
Born in New York, Noma attended Dartmouth College, an Ivy League school in the picturesque town of Hanover, New Hampshire.
A math major, Noma worked as a systems programmer in the school’s computer center. Several of his friends were psychology majors. They told Noma that he should meet one of their professors—John C. Baird—who taught psychology and brain sciences and was looking for a programmer to help with a project.
Noma began working with Baird in the field of psychophysics. They wrote programs that analyzed psychological data, such as the scenario posed at the beginning of this article. They even wrote a book together, Fundamentals of Scaling and Psychophysics, published in 1978.
“You collect all this data and you have to analyze it,” Noma says. “How is this person’s bias different from that person’s?” Take in large amounts of both structured and unstructured data, crunch it down, make it so it’s readable, analyze it—sound familiar?
Noma moved on to the University of Michigan to get master’s degrees in both mathematics and psychology; in 1982, he became a PhD in mathematical psychology, with a focus in behavioral finance. “A lot of what we see in the stock market is momentum; it’s psychology," he says. "There was a desire to have tools to figure out what movements meant when people got excited,” he says.
Wall Street Roundup
After serving for three years as an assistant professor in Rutgers University’s psychology department, and after a brief stint as a statistical modeler at Booz Allen Hamilton, Noma finally made his foray into banking. In 1987, he joined Chase Manhattan Bank as a mortgage banker where he performed mathematical analysis of options and statistical modeling of prepayment rates used in the pricing of mortgage-backed securities (MBSs).
He also modeled term-rate structures using time-series analysis. He moved on to First Boston in 1990, where he modeled default and repayment programs for manufactured housing contracts. A year later, he hopped over to the buy side and joined Salomon Brothers Asset Management as a portfolio analyst, where he advised on the controlled downsizing of a $5.1 billion portfolio of various derivative mortgage-backed securities held by Franklin Savings Association.
Two years later, it was back to the sell side, where he joined Deutsche Bank as a director overseeing market risk management on fixed-income and mortgage derivatives. There, he developed a methodology for calculating the value-at-risk (VaR) of US government bonds, MBSs, secondary corporate trading, debt capital markets, repos, and the bank’s treasury liquidity portfolio. In 2000, it was on to Merrill Lynch, where he was the risk manager for fixed-income funds, with a focus on corporate investment grade, mortgage and money-market funds.
“I would have loved to have been a banker forever. Granted, I had no life because it was a hundred-hour week where you don’t do anything but work. But it was fun," he says, with a chuckle. "I don’t know how people do it for 30 years, because you can’t avoid getting burned out, but for that moment in time, I had a wonderful time.”
After 16 years, Noma felt he had accumulated enough experience on the sell side; he next wanted to learn the structure of the hedge fund world.
“The main reasons why I left Merrill Lynch were because, one, hedge funds were becoming hot and I desperately wanted to get into the business; and two, I felt I was senior enough to be able to run a department and even create a department,” Noma says. “That’s what Asset Alliance offered me.”
In 2002, Asset Alliance had been hit hard after Beacon Hill—in which Asset Alliance had a 50 percent investment—lost $400 million in failed US interest-rate positions. It needed to improve its risk management practices, and Noma was named the firm’s chief risk officer (CRO), overseeing all multi- and single-manager offerings.
While he was used to managing risk, he also wanted to manage a book. So in addition to his CRO duties, he became a portfolio manager for a fund of commodity trading advisor (CTA) funds, which contained about 20 managed futures, commodity and macro programs, including systematic and discretionary strategies. In 2007, the fund of funds had a net return of 10 percent, while in 2009 its return was 5 percent. It was at this time that he decided to go it alone, founding Garrett Asset Management.
While Noma still continued as an advisor with Asset Alliance, he wanted to test what he had learned as an active manager and build his own strategies, which led him to found Garrett Asset Management in 2008 and launch the fund in 2009.
As noted earlier, the fund managed just under $5 million at its peak, covering futures, ETFs and currencies. Right now, the firm is in the midst of a transition. All investor monies have been returned, and Noma is winding down the active part of his book, while Manvae moved on to work for communications vendor ViaSat in 2014.
Noma says he is not sure whether he will close Garrett and start a new fund, or simply launch a new portfolio. The plan, though, is to stay in the game as a hedge fund manager. What will change with this next iteration, though, is the use of artificial intelligence (AI). Previously, Garrett was based on classic technical strategies; the thought going forward is to deploy genetic algorithms, random forest and deep-learning techniques.
“Over the last several years, the markets have become more correlated,” he says. “The correlation means that it’s harder to get diversification in some sense, but it also means that there might be a greater opportunity to get signals. You can use the fact that multiple markets are doing the same thing as a magnifier of what the signals are.”
In the field of AI, there are several different techniques that can be deployed, such as neural networks, genetic algorithms and decision trees. It’s the latter of the three that has seen a fair amount of traction on Wall Street, according to Noma.
Decision trees and random forests help a trader to understand the basic processes of the strategy. Other AI techniques, though, need more development because they are more “black box,” making it harder to explain the reasons why trading decisions are being made. But for trading and asset allocation, there is potential that can help a small hedge fund act more nimbly.
“The market has also become a bit scarier in terms of the half-life of strategies,” Noma continues. “The idea is to figure out some way to build in greater automation so that a small staff can turn over these strategies and have a better, more automated system to develop strategies, screen them, and implement them quickly and responsibly. There’s a lot to be explored with AI and a lot that remains to be seen as to which areas of finance present the greatest opportunities.”
It’s the kind of thing that piques one’s curiosity.
There’s something else to keep in mind when talking about Noma: Despite the fact that he is 65, he still held three different jobs at the start of 2016. In addition to serving as an advisor for Asset Alliance—which he stopped doing a few months ago—and managing his book at Garrett, in 2012 he created a company called Claritica, which creates a variety of products, such as mobile apps, IoT devices for weight training, and helmet-mounted accelerometers.
Noma is a born tinkerer: He is the kind of guy who buys his family an Amazon Echo for Christmas, but hacks it and installs an app that allows you to talk to it and have it turn the lights on and off on the Christmas tree. “I’m just a tinkerer,” he says. “Some things don’t have a financial benefit; at this point it’s about having fun.”
When asked about what advice he would give a young programmer who does not want to spend 20 years at a single bank—a programmer who dreams of one day starting their own hedge fund and deploying their own strategies—Noma doesn’t have anything groundbreaking to say: “The main skill is to understand that with every job you have, there’s a great opportunity to learn something new, and to exploit it.”
To him, the answer is simple: Be curious, work hard, and don’t be afraid to follow those curiosities. Tinker and have fun.
And when will this decades-long odyssey finally wind down? The answer appears to be not anytime soon. Noma will fire up his AI strategy sometime down the line, perhaps in eight months to a year, and when asked about when he plans to retire, it’s the only question of our interview that he appears uncomfortable fielding: “It’s unclear to me at this point what that means,” he says.
A tinkerer never stops tinkering.
Elliot Noma Fundamental Data
Name: Elliot Noma
Education: Dartmouth, A.B., Mathematics (1972); University of Michigan, M.A., Mathematics (1979), M.A., Psychology (1979), and Ph.D., Mathematical Psychology (1982)
Kids: A son who is 18 and will follow in his dad’s footsteps and attend the University of Michigan
Hobbies: Biking, because it allows him to disconnect from his phone and the computer. He used to be an avid runner and skier.
Waters Wavelength Podcast Episode 75: An Update on the Julia Programming Language; AI & Alternative Data; Digital Currencies
Julia Computing's Viral Shah talks about the programming language he helped create and what's ahead for it. Then James and Anthony talk about the pairing of AI & alternative data, digital currencies, and Game of Thrones.Subscribe to Weekly Wrap emails