Can we salvage the idea that factor models reflect risk? Perhaps. McLean and Pontiff find that only 58 percent of the returns associated with factors disappear after publication -- that still leaves 42 percent. This could reflect underlying risk associated with the factors. Alternatively, it could be the result of investors’ inability to fully exploit mispricings -- an effect behavioral finance researchers call "limits to arbitrage." Mispriced stocks could be difficult to borrow and sell short, for example. So although McLean and Pontiff's paper shifts the debate in favor of the behavioral side, it doesn’t resolve it.
Still, the existence of mispricings that are so large and persistent that academics -- rather than investors -- are the first to discover them deals a strong blow to the efficient markets theory.