Irreconcilable Differences or Calculated Irascibility?

“Smart-Beta War Rages On as Cliff Asness Slams Arnott’s Paper” or so says Bloomberg. But perhaps the ideas of war and street smack downs are, dare we say, media hyperbole? AQR’s Asness claims that factors can’t be timed (mostly). Rob Arnott of Research Affiliates says certain factor valuations have become stretched. They’re both right, and the two, who have collaborated on research in the past, would probably agree with each other behind closed doors. But where’s the fun in that? Controversy brings attention, and as they say, all PR is good PR.

Why do we suspect that they probably agree more than not on this issue? Both Research Affiliates and AQR calculate and publish capital market assumptions, or return forecasts for various asset classes, and they both directly acknowledge (as do most who formulate CMAs, ourselves included) that starting point, in terms of valuation, is highly predictive of future returns, albeit silent as to timing. So yes, factors can be overvalued, and no, that doesn’t mean that investors can use relative valuation as a short term timing tool for personal gain; all asset types can remain greatly over or undervalued for extended periods of time.

Factor exposures have always been sought out by managers, perhaps not in such a direct fashion as they are today, but all investment strategies – systematic and discretionary – attempt to isolate characteristics believed to offer outsized payoffs. Our take is that factors do indeed have distinct payoffs, that those factors can exhibit momentum, and that momentum can be identified and exploited, so long as a manager protects against the inevitability of factor mean reversion. So we’ll let the Titans of Finance face off for the show of it; we’ll continue to exploit the middle ground.