How Covid-19 shocks revealed machine-based hedge funds’ “Achilles’ heel”
This year’s Covid-19 stock market upheaval has revealed an “Achilles’ heel” in quantitative hedge fund models that use traditional factors such as momentum and value, as machine-based strategies have struggled against “unique new drivers of stock returns that don’t fit the academic models”, says Man FRM.
2020 proved to be a “blow to the head” for factor-based quantitative equity analysis, Man FRM observed in its ‘Early View’ commentary on Wednesday.
Quant equity hedge funds which trade traditional factors have struggled amid the dominance of a new “Covid factor” this year, with “near-consistent disappointment” in the form of “losses at the start of the year, losses in the market sell-off, losses through the summer and, to compound matters, losses in November.”
The note, written by Keith Haydon, CIO of Man Solutions and Adam Singleton, head of investment solutions at Man FRM, explored how shorter-term market upheaval – such as the dot-com crash of the early 2000s and the 2008 Global Financial Crisis – can often render factor labels like value and momentum redundant as they exhibit negative correlation during shocks.
“Trying to force stock behaviour into one box or another can be a good idea in normal times, as it lets models identify distinct baskets of stocks out of a large universe that best explain each factor,” they wrote.
“But this falls down when the majority of stocks are being driven by a new influence, as this narrows the breadth of the equity market drivers and makes it harder to find stocks which are uniquely value or momentum, or anything else for that matter.”
Although losses in value factors outweighed smaller gains in momentum for much of the pandemic, and November’s factor rotation saw losses from momentum overwhelm value gains, Man FRM – the fund of funds unit of London-headquartered hedge fund giant Man Group - suggested this analysis may ultimately be a mischaracterisation of market behaviour.
“Trying to compel new market forces to submit to the old models of market behaviour has led to chaos for quantitative equity market hedge funds – no more so than in November, where on the vaccine announcement days the Covid factor was the only game in town.”
While value stocks surged and momentum names tumbled following Pfizer’s vaccine breakthrough last month, Covid-19 remains a new “nascent influence” on markets, making it “impossible” to calibrate equity risk models particularly well.
“Perhaps a better way to describe it was that the Covid factor went sharply into reverse, and the more familiar – but wholly inadequate – labels convulsed wildly in its wake,” Haydon and Singleton wrote.
“If nothing else, it’s clear that 2020 represents an Achilles’ heel in the hedge fund machines, namely a real failure to grapple with unique new drivers of stock returns that don’t fit the academic models. Even the newer ‘machine learning’ approaches felt somewhat impotent, since you need to learn from something, and probably rack up losses while you’re learning.
“It seems to us that, even if it is only to point the machines in the right direction, we may still need humans in this game for some time yet.”