March 24, 2021
Factor investing has long faced criticisms of data mining, and more recently faced another criticism – some backtests might never have been right to begin with. A growing body of mostly well-done papers examine these issues, generally concluding that factors have been disappointing since their “discovery.” We’ve long addressed these concerns through robust out-of-sample evidence and a compelling theory for why a factor should work. What we’ve lacked, until now, is a formal test. My colleagues’ new paper tests brilliantly, what we have argued, largely anecdotally, for years. Their results are rather startlingly (even to me) positive for the field in general.