It was the kind of volatility event that fundamental investors feared might trigger indiscriminate selling in black box trading strategies.
On May 17, the S&P 500 plunged in the wake of news reports that US president Donald Trump pressured former FBI director James Comey to drop his investigation into the activities of Michael Flynn, the US national security adviser.
But the subsequent 1.8% fall in the US equities market – a 2.3 standard deviation event – barely ruffled the feathers of systematic trend followers who have been accused of being ticking time bombs in the event of a volatility spike.
In fact, some analysts now believe the fear of co-ordinated unwinds in systematic trading strategies run by commodity trading advisers (CTAs) is a greater risk to the market than a CTA sell-off itself, because it could trigger a spiral of unwinds from discretionary traders.
“If you’re a fundamental or discretionary manager and you have people telling you black boxes are unwinding billions of dollars, that sounds scary. But the bigger risk is the fear itself,” says Chintan Kotecha, equity derivatives analyst at Bank of America Merrill Lynch.
In the extreme case that CTAs should unwind their equity exposure in a week-long scenario of high-sigma equity market declines Kotecha’s team estimates this would equate to just 6% of expected equity futures trading volume.
“Conceptually, CTAs are potentially selling at a time when the markets are declining, and that would add extra pressure, but even in their most extreme selling situation, based on the assets we believe they have, it wouldn’t be a substantial percentage of total market volumes,” Kotecha says.
That’s despite BAML in the past having raised concerns about CTA selling pressure, though others on the sell side have taken a more strident position.
Marko Kolanovic, JP Morgan’s global head of quantitative and derivatives research – who declined to be interviewed for this story – was the first to warn that black box strategies such as CTAs and risk parity funds could wipe $100 billion off the stock market in a stress event.
The market should be able to handle [a CTA sell-off], but what it can’t handle is if other people are selling because of this
Chintan Kotecha, BAML
CTAs say the fears are overblown. “When you look at what portion of the market CTAs make up, you have to be somewhat realistic about how much swing we can have when things unwind,” says a source at one of the largest CTAs.
CTAs lost 1.1% on May 17, according to data from Societe Generale’s CTA index. The strategy lost another 43 basis points the following day, and was still slightly down by week-end but there was no downward spiral (see figure 1).
BAML’s Kotecha says: “The market should be able to handle [a CTA sell-off], but what it can’t handle is if other people are selling because of this. In a stress event, when you have this added fear factor, it could create a feedback loop which is the biggest concern.”
CTAs’ exposure to equities is currently close to historic records due to positive price trends, according to BAML’s research, which estimates CTAs are currently long somewhere between $70 billion and $175 billion in global equities.
However, in a note published just weeks before Comey was fired, the bank predicted it would take five successive days of two-sigma declines to force CTAs to unwind equity positions entirely.
CTAs say it would take an even longer-lasting slump. “What needs to happen is a persistent change in trend,” says Tim Pickering, chief information officer at Auspice Capital Advisors, a CTA based in Calgary, Canada. “The equity trend has largely remained up for some time. It would take a significant shift to change the accepted trend from up to down.”
Even then, CTAs say the diversification of time horizons in their strategies means not all funds would be affected in the same way. “It may impact short-term traders responding to three to 10 day trends but not medium to long term CTA trend traders that make up the bulk of the sector and exposure,” Pickering says.
It’s hard to imagine a massive sell-off due to a five-day reversal
Marat Molyboga, Efficient Capital Management
James Dailey, chief executive officer of Dunn Capital Management, estimates approximately two-thirds of AUM in managed futures is made up of long-term trend followers.
“As far as a one-day move being material enough to trigger a sell-off, that could be true for short-term managers, but around 55%–65% of AUM in managed futures is made up of long-term trend followers, so a one-day event would trigger people trimming positions because of a volatility spike but it’s not indicative of a reversal of a trend,” he says.
Marat Molyboga, chief risk officer and director of research at Efficient Capital Management, says both the duration and the magnitude of the move are important. “A lot of CTAs use signals that are similar to 12-month time-series momentum. There could be some reversals in positioning of short-term managers but it’s hard to imagine a massive sell-off due to a five-day reversal,” he says.
CTAs have been defending themselves since Kolanovic published his damning research report blaming three types of “price insensitive” trading strategies for the US equities flash crash on August 24, 2015.
According to that report, if equity indices fell 10%, CTAs, risk parity and volatility managed strategies might need to sell as much as $100 billion of equity exposure. “These types of price insensitive flows are starting to materialise, and our goal is to estimate their size and timing,” Kolanovic wrote at the time.
In April, prominent hedge fund manager Paul Tudor Jones said systematic trading strategies would act as “the hammer on the downside” when volatility returned to the equity markets.
Many systematic managers say the combined equity exposure of risk parity portfolios – around $50 billion – and CTAs is insufficient to impact the wider market.
“For a variety of reasons, [risk parity’s detractors] don’t have the numbers right. We know the actual sizes and the estimates are just way off,” says Michael Mendelson, portfolio manager of AQR’s risk parity strategies.