Evolving alpha: Long-running quant fund Millburn Ridgefield’s 50-year strategy for success

Quant trading

Over the course of its remarkable history, global quantitative investment manager Millburn Ridgefield Corporation has evolved through several iterations – from starting out as one of the early pioneers of traditional trend-following in the 1970s to becoming a major innovator and practitioner of machine learning quantitative investment approaches during the past decade. Now, with the launch of a new strategy, Millburn aims to build on its success by including exposure to global, Chinese and thematic commodity markets.

The firm, which this year celebrates its 50th year in business, is a systematic manager, running some USD6.6 billion in assets – including those managed by a joint venture in China – with 57 employees, nearly 40 per cent of which are involved with technology-focused research and development. While the firm cut its teeth in futures trading, today it has developed a capability in futures, forwards, and securities.

“It’s been interesting to see the firm evolve in the way it has,” Barry Goodman, co-CEO and executive director of trading, tells Hedgeweek.

“Whereas in the 90s and early 2000s, you would have seen us focusing on longer-term price momentum and being relatively slow in our trading frequency, the current approach is very nimble.”

Goodman – who has been with the firm for nearly 40 years – explains how the systems are designed to evolve, making fine-tuned adjustments throughout the day, in real-time as the data flows in.

“We may still find ourselves in a long or short position for a period of time, but as new data flows into the suite of models, like a spike in short term momentum, or a flattening of volatility, or maybe the release of some new inventory numbers, or even increasing proximity to the next Fed meeting, the models have the ability to interpret this and change that signal fairly rapidly.

“But whether the signal changes or not, and to what extent, is now context-dependent, rather than based on simple rules or hypotheses as it may have been in the past.”

A data-driven edge

Today, Millburn applies this approach in the management of a broad range of primarily long/short strategies, with its models designed to be particularly agile in a diverse selection of markets and regions in various timeframes.

“One of our key cornerstones is diversification. We’re not global macro traders, we’re data-driven in our approach – so we always try to provide very diversified access with the view that as many of these markets are going to move and be volatile, we can trade on that volatility,” Goodman adds.

Reflecting on the firm’s more recent evolution, he says that as investors grew increasingly comfortable with complex algorithms and machine learning-based strategies, Millburn Ridgefield was quick to capitalise on this shift in sentiment.

While there was always a certain class of investor who believed in quant, and were willing to sacrifice a degree of transparency in exchange for the level of alpha that they believed quants could generate, the increased importance and perceived value of data within the alternative investments sphere during the past decade has accelerated this change.

“We believe that there’s an increasing recognition amongst even traditional investors that data is key,” he continues. “It’s become accepted that data can provide an edge, even on the discretionary side, with more people now focused on the data that they’re using and evaluating.”

Against that backdrop, he describes the firm’s systematic trading strategies as what he believes to be a “common sense approach” to making use of data and computing power.

“For us, even though we rely on a highly technical and very sophisticated machine learning approach to finding patterns in historical data, we try to limit the problem for the machine, providing it with a lot of data, but making sure the data has, in our view, at least some economic rationale and a reasonable chance of influencing price moves. So for us it is really the combination of our experience with the power of the machine,” Goodman observes.

“We let the machine do the difficult work of finding those complex patterns that humans simply could not find. But we set up the framework with the goal of enabling us to explain to investors what was driving the signal at any particular time. We try and make sure people understand our process, and get comfortable with the process.”

Building on this theme, he continues: “Based on history, we may sometimes look like a trend-follower, meaning we’ll be following trends and riding those trends out. But at other times, we appear more like a short-term trader, taking what look like mean-reverting positions or holding positions for perhaps just a few days, or even a few hours.”

Reaping rewards

It’s an approach that has ultimately reaped rewards for the firm’s flagship diversified offerings, but also for its sector-specific strategies.

The diversified offerings comprise two identical quant strategies – the Diversified Program, whose track record stretches back to 1977, and the Multi-Markets Program. Each one trades around 100 liquid futures and forward markets globally, spanning equities, fixed income, currencies and commodities. Together, the Diversified and Multi-Markets strategies manage approximately USD4.5 billion in assets, the bulk of Millburn Ridgefield’s AUM. Performance-wise, the Diversified and Multi-Markets strategies have generated net returns of more than 11 per cent so far this year, with correct calls on stock indices and energy sectors helping to drive returns.

The firm has also established a formidable presence in more specialist sector- and region-focused systematic strategies.

The first of these is the Millburn Commodity Program, trading a range of commodities futures. This strategy, which has been closed to new capital since 2018 as a result of strong demand, has climbed more than 10 per cent net so far in 2021, with much of those returns coming from gains in energy, as well as grains and precious and industrial metals. Notably, the strategy delivered on the promise of diversification in 2020, returning nearly 10.9 per cent net during the Covid-driven maelstrom in March last year.

Meanwhile, the China Futures Program, which Millburn developed with a local on-shore partner through a joint venture, provides access to a range of unique markets to local Chinese investors. After returning more than 46 per cent net in 2020, its returns have been comparatively muted this year, up some 4 per cent net.

In June, the firm unveiled its long/short Resource Opportunities strategy. Like the Commodity Program, its capacity is limited – the Resource Opportunities strategy is thought to have an initial limit of between USD350-500 million, although with continued research the hope is this will grow.

“Providing access to truly global commodity futures markets is exciting, but the strategy also seeks to build thematic trades across disruptive global sectors and markets such as wind, solar, water, rare earth metals, lithium and more,” Goodman says, describing it as a “next generation commodity strategy.”

Navigating volatility

As the Covid-19 pandemic has upended global economies over the past 18 months, and the fallout from various central bank responses still carries potentially far-reaching responses, Goodman maintains that any potential movements towards increased volatility can be a positive for Millburn Ridgefield’s strategies.

“There are many things going on – decarbonisation trends in commodities; supply chain disruptions, which create supply/demand imbalances; how the actions of central banks can affect inflation – all of which we believe can be supportive of the way we trade,” he says.

“We’re fully long/short in the majority of our approaches. We have no bias to whether we want to have a long view or a short view in a market, and we believe we are in a position to take advantage of that and help investors navigate this volatility.

“Our strategy is not static. The models that we build and the data that we use are always evolving. The models that we built eight years ago and the ways they interpreted markets and data might have a very different view of the world in 2021. Certain drivers of the markets may be much more important now than they were back then – and other drivers may take a back seat now. That’s why we continually update and refresh the models.”