Shifting paradigms: How hedge funds are navigating the data “explosion”
This year’s hedgeweekLIVE Technology Summit concluded with deep-level discussions of how hedge fund managers running a range of strategies are grappling with the rapid pace of change in data science and how the “explosion” in data volume can be managed more effectively.
Kicking off day two, the data management and analytics panel addressed how the increased volume of data is reshaping the investment process. Speakers also mulled the value of datasets in generating returns over the longer-term, and whether such alpha is ultimately ephemeral and short-term.
Jatin Dewanwala, head of research at Hildene Capital Management, noted that in his firm’s area of focus – structured credit and CLOs – data moves at a “much slower velocity.”
“We still have challenges, and they become more relevant when paradigm shifts take place,” he explained. For instance, the Covid-19 crisis saw disposable incomes rise along with unemployment levels as a result of government stimulus, which “broke the correlation” and brought major implications for datasets in credit assets linked to consumer receivables.
“So when paradigm shifts happen, it’s very important to be close to the ground and understand what’s going on, and at every single stage be aware of what your model may be capturing and what it may be missing. For us, it’s very important that the analytics keep evolving to capture whatever topics are in focus, and that we enhance our analytics accordingly.”
Rani Piputri, head of automated intelligence investing at NN Investment Partners, and Mani Mahjouri, CEO/CIO at quant-based equity market neutral firm Blueshift Asset Management, further explored a recurring theme at this year’s event – the convergence of discretionary and quantitative investing.
Piputri suggested this balance between disciplines will become “more and more prevalent”, and will offer market participants a “different way of looking at things.”
She acknowledged how certain data which is relevant for equity investors is usually very different from fixed income data, making uniform, unified data systems difficult to create. Yet she also suggested that investors’ and managers’ ability to scrutinize and manipulate data has improved, and can be better tailored towards their individual needs.
Mahjouri meanwhile flagged the “explosion” in the availability of data, and the multitude of applications being used to slice the data in new ways.
“There’s more data now than ever before and it’s growing at an unprecedented and rapid rate,” Mahjouri said. “There’s also a spectrum of alpha that hasn’t really been covered. A discretionary manager holding a handful of positions needs to have certain level of conviction before entering a position; a quant manager may have a lot of positions with a lot less discretion. Somewhere in the middle there is a lot of risk premium and alpha that has yet to be discovered and utilised.”
Later in the afternoon, the concluding panel of this year’s event travelled deeper into data science and research processes, with speakers weighing up the ways that hedge funds can better optimise information and calibrate their systems to offset biases.
“Every dataset has its own biases,” Qaisar Hasan, portfolio manager, Lombard Odier Investment Managers said. “Different datasets work in some situations for some names in certain markets, but won’t work for others.” Against that backdrop, Hasan outlined some of his firm’s initiatives centred around combining different datasets together to create “synthetic hybrid signals” which “we find much more powerful” than different datasets that sit in their own silos.
Julien Messias, co-founder and portfolio manager at Quantology Capital Management, observed how old investment strategies and research papers stretching back 50 years are being revamped and remodelled using state-of-the-art algorithms, with technology providing a new edge and, crucially, good returns.
“Algorithms are the motor engine and the data is the gas,” Messias said. “Without the appropriate gas, the car won’t move. You can have a very powerful algorithm, but if you don’t have the right data, you won’t go anywhere.”
Acknowledging systematic strategies’ mixed performances in recent years, Angana Jacob, head of product management at SIGTech, believes that as a result of the pandemic, there is a growing sophistication in quant techniques, with a focus on mitigating the sector’s traditional drawbacks.
“So there’s less calibration on historical data and more of a focus on making quant strategies more adaptive and responsive to regime changes,” she explained. “There is also a much bigger focus on risk management, additional stress testing, and more simulations to understand how strategies might behave in unexpected conditions – like a pandemic, or climate risk, or less extreme situations like sudden curve steepening or inflation.”