Momentum stocks have taken a hit recently, but those who say that indicates momentum investing is not a profitable long term approach are dead wrong, say the authors of a new paper.
In the paper, Clifford Asness, Andrea Frazzini, and Ronen Israel of AQR Capital Management and Tobias Moskowitz of Chicago Booth seek to dispel what they say are 10 myths about momentum investing, and Morningstar’s John Rekenthaler provides a nice summary. For example, to those who say returns from momentum strategies are too small and sporadic, they counter, “[Momentum] is present in U.S. stocks over very long time periods and following its academic ‘discovery’ in the early 1990s, has been shown to be robust out-of-sample (an important exercise we will repeat here), in the individual stocks of other countries, for stock markets, and for completely different asset classes, such as bond markets, currencies, commodities, and others.”
They also address the contention that momentum strategies are too volatile. “As with any factor, momentum does not make money all the time and occasionally suffers large losses, and historically this has been somewhat worse for stand-alone momentum than the other factors discussed here,” they say. “Spring 2009 was one of those times. But, while more extreme, this isn’t unlike other factors. There is a saying at the University of Chicago, ‘the plural of ‘anecdote’ is not ‘data’.’ Neither 2009 for momentum nor 1999 for value are indicative of the overall health and strength of these strategies.”
The authors also discuss the theory behind momentum investing — a theory critics contend doesn’t exist. “The idea is that if the momentum premium is really as large and robust as we show it to be, then it must be due to a market inefficiency and therefore (and here’s where the religion comes in) it can’t be real, as the markets are obviously perfectly efficient,” they write. “While we believe that risk-based efficient market explanations play an important role in all of these factors’ returns [i.e., size, value, momentum], we also believe there is a role in each, perhaps at different degrees, for behavioral explanations.”