Who is better at factor investing: Cliff Asness or Rob Arnott?
The rivals and founders of AQR Capital Management and Research Affiliates, respectively, seem determined to duel it out. Their latest public comments over data mining – the bad habit of “searching the data to find in-sample patterns in returns that are not real but random, and then believing you’ve found truth,” according to Asness – are no exception.
This most recent round in the ongoing feud began last week when Bloomberg published an article on data mining in which Arnott is quoted describing Asness as “insufficiently skeptical about the pervasiveness of data mining and its impact even in the factors he uses.”
Asness – never one to shy away from conflict – responded in blog form on Wednesday, bristling at what he interpreted as Arnott labeling him a data-miner. While acknowledging that overreacting is “kind of [his] go-to move,” Asness called Arnott’s comments both “provably false” and “particularly hypocritical.”
“At AQR we pride ourselves on minimizing data mining,” he wrote. “Nobody in our field is perfect on this front but we’ve had the discipline to walk away from good-looking factors we don’t trust.”
According to Asness, AQR and Research Affiliates largely endorse the same investment factors: value, low risk, and momentum. So if AQR’s factors are data mined, he argued, so are those used by Research Affiliates.
“That he’d accuse us of being ‘insufficiently skeptical’ about the dangers of data mining isn’t just at odds with our long history of the exact opposite, but bat**** crazy when it’s mostly the stuff he believes in too,” Asness wrote.
Asness and Arnott’s dispute over data mining is not new: Arnott has previously accused the factor investing community of significant data mining – specifically, “a rather extreme form of data mining” that has made many academics and practitioners blind to what he sees as rising valuations. In other words, he believes some factors have become too expensive – and it follows that investors should time their exposures to factors to achieve the best results.
Asness, for his part, has repeatedly critiqued Arnott’s research, dismissing the notion that AQR’s factors are both data mined and overvalued. Most recently, he and three AQR colleagues published a paper asserting that factor timing was far too difficult to be carried out with economically meaningful results – and that current valuations were nothing to worry about anyway.
To date, Arnott has never formally responded to Asness’s criticisms.
