Wall Street’s Most Famous Quants Fed Up With JPMorgan SoothsayerBy
Risk parity, trend-following strategies used as scapegoat: AQR
Size of industry overestimated in bank research, firm says
Don’t listen to Marko Kolanovic.
Or anybody else who tells you quant managers regularly whip up bouts of pain and suffering for stock investors. Funds with programs that follow trends and sell like robots are getting smaller and simply aren’t big enough to overwhelm the $24 trillion U.S. equity market.
That view comes courtesy of AQR Capital Management which, it should be noted, is far from an unbiased observer, given the $172 billion investment house pioneered many of the strategies coming up for vilification. AQR’s quants say they’ve had it after listening to more than a year’s worth of hectoring from analysts who blame managed futures and risk-parity strategies for everything from August 2015’s China meltdown to the post-Brexit plunge.
“The analysis is inaccurate,” said Brian Hurst, a principal and portfolio manager for AQR’s managed futures and risk-parity strategies. “They are using oversimplified models with bad inputs.”
However self-serving, AQR’s complaints amount to more than quibbles among math geeks. The past year has seen an ascendancy in the type of analysis the investment firm is describing, efforts to predict the impact of automated traders by estimating how much money they control and theorizing about their reactions to things like volatility.
As AQR sees it, computer-driven funds are just scapegoats in a doomed quest to explain jarring market movements. Topping the enemies list is JPMorgan Chase & Co.’s Kolanovic, viewed by many as a gnomic visionary after calling an equity rout in the summer of 2015, blaming forced selling by automated funds.
Those kinds of calls have earned Kolanovic a perennial slot atop Institutional Investor’s research rankings and one of the most devoted followings on Wall Street. His forecasts around the August 2015 crash, when the S&P 500 plunged 11 percent in six days, was viewed by fans as scarily prescient and one of the only coherent explanations of that event.
Kolanovic declined to comment, according to Amanda Smith, a spokeswoman for JPMorgan.
AQR’s quants aren’t in the cheering section. They wonder where the data is coming from. Take the stretch following Britain’s vote to quit the European Union in June. Kolanovic posited that deleveraging by automated funds would spur as much as $40 billion in selling by commodity trading advisers and $30 billion by funds using the risk-parity framework.
“These figures could be higher if realized volatility continues to increase,” he warned.
AQR says estimates like that grossly exaggerate those funds’ sway. For example: a sizable chunk of money run by CTAs is in managed-futures and trend-following strategies, it says, with managed futures’ assets dropping to about $135 billion from $210 billion in 2008, data from eVestment show. Compared to the $300 billion to $400 billion that trade in S&P 500 futures daily, they’re not big enough to stampede the market.
“We still have to fight the myth because people come to us and say, ‘The managed futures category has gotten so big, aren’t you guys really worried?’” said Yao Hua Ooi, a principal on AQR’s global asset allocation team. “We have to explain that actually, no, it hasn’t. That’s a misperception.”
Getting a precise handle on who’s right is hard because of the squishiness of definitions in the quant space. Kolanovic writes about CTA funds, a larger universe than just managed futures. CTAs do so many different things that pinning down their moves as a group is even more difficult, according to AQR.
Often, researchers conflate CTA and risk parity funds, though they buy and sell based on different signals. Risk parity doesn’t chase trends, but rather is a strategy for partitioning a portfolio according to volatility averages over a medium-to-long time horizon. But whether it’s risk parity or CTAs, there’s a large diversity in signals among different quant funds, making any attempt to pin down overall positioning futile, said Hurst.
“Systematic strategies are often lumped together into one group, when in reality there is a vast array of different quantitative or systematic strategies that are pursued -- many of which have little to do with one another,” he said. “The variety of strategies that exist in the systematic space is often overlooked.”
Compounding the issue, the analysis often gets positioning and timing wrong, according to AQR. Take Sept. 9, when the S&P 500 plunged the most since Brexit after central bankers signaled reluctance to extend stimulus. Equity analysts at Bank of America Corp., led by Chintan Kotecha, estimated that quantitative-driven funds could contribute $40 billion of near-term selling pressure, triggered by the price gyrations.
But one day’s rout will almost never exact enough pain to force those funds into that big a rebalancing -- most CTAs follow trends on a one-month to one-year time horizon, AQR holds. Even on the shortest end, the strategies aren’t hyper reactive.
AQR’s own risk parity strategy moves slowly, and sold only a small fraction of what Bank of America estimated in September, said AQR’s Michael Mendelson. In fact, there hasn’t been a single instance this year where the fund saw significant deleveraging, he added.
“Our risk model can’t move around as fast as that or we’d spend everything we’ve got on transaction costs,” said Mendelson, portfolio manager of the firm’s risk-parity strategy. “It’s just not the real world.”
AQR’s criticism of the research isn’t altogether new. In fact, other large banks have maligned notes from Kolanovic and the like. In August 2015, responding to JPMorgan’s research, Societe General SA said the claim that robotic selling could ignite crashes was an oversimplification. Systematic strategies could not be drivers, only amplifiers of selloffs, SocGen’s Kokou Agbo-Bloua said at the time.
Even within the ranks at JPMorgan there’s been push back. The bank’s flow-and-liquidity team, led by Nikolaos Panigirtzoglou, said that traditional managers had sold more than quantitative players this fall.
Bank researchers aren’t the only ones guilty of quant-blaming. Hedge fund managers have also used the robots as a scapegoat. Last year, Martin Taylor of Nevsky Capital closed his 15-year-old fund, lamenting the distorting influence of computer traders.
“It’s ‘the dog ate my homework,’” said Hurst. “The reality is market participants don’t have any insight into our exact models and trading strategies.”
— With assistance by Oliver Renick