Quant traders at Man Group Plc are betting that unlocking the secrets of private markets will give them an edge in trading public stocks.
Now worth $13 trillion, unlisted assets are snowballing while the number of publicly-traded firms has steadily shrunk. That’s made them into a force too big to ignore, according to Man Numeric, the quant arm of the world’s largest public hedge fund.
All in, the firm expects private market data could one day represent 10% or more of the trading signals for some of its investment strategies. Thanks to advances in artificial intelligence like so-called natural language processing, insights from private markets – like revenue data for consumer services – are now helping to inform Man Group’s systematic trading.
“Private market companies have much closer proximity to the real pulse of the economy,” said Ori Ben-Akiva, director of portfolio management at Man Numeric. “They capture changes in the underlying economy and reflect that in their fundamentals faster than they get reflected into public market companies.”
The firm contends it’s able to capture a leading indicator for public equity trading from data covering a universe of over 1 million companies that don’t trade at all.
Even as they’ve grown, private markets have mostly resisted efforts to be measured, valued or traded, much less become the basis for quantitative trading strategies. That hasn’t stopped a growing number of hedge funds from trying to channel alternative data into models that used to rely mostly on equity valuation and price momentum.
Similar insights will be used to underpin a new exchange-traded fund Man Group is readying with KraneShares that will buy listed small and mid-cap stocks.
The urge to unlock the private market black box is already driving a race between ETF giants who want to make them into funds for the masses, while BlackRock Inc. heralds its ambition to “index the private markets.”
Data in private markets is patchy because private companies don’t have the same mandatory disclosure requirements as listed peers, making it harder to feed to algorithmic trades.
But recent advances in NLP have made it possible to access a lot more details about private companies — and by extension make inferences about public ones.
Man Numeric expects to double the weight it places on private market data-driven models in the next 12 months. The quants use NLP to extract top-line revenue figures from public records including company reports, social media, news outlets and government agencies.
Private equity deals are another valuable source of information. Market-moving disclosures of patents and hiring plans can be contained in deal documents.
“There were more opportunities to model things at a more stock specific level,” Man Numeric’s Ben-Akiva said. “It’s an evolution in the available data.”
Private companies represent the broadest swathe of corporate America, outnumbering public ones by more than a thousand times, according to Man Numeric research. They vary widely, with annual revenues from as little as $1 million to more than $100 billion.
A preponderance of business services and industrials are in private hands, according to Man Numeric research. That contrasts with the public market’s skew toward financial and healthcare sectors.
Academic research corroborates Man’s findings. A July paper authored by Paolo Barucca and Flaviano Morone at University College London and New York University found that private equity transactions predict public market behavior with about 70% accuracy, especially for consumer services and communications.
A message from Advisor Perspectives and VettaFi: To learn more about this and other topics, check out ourvideos.
Bloomberg News provided this article. For more articles like this please visit
bloomberg.com.
