Here's his latest survey:
We thought returns were uncorrelated over time, so variation in price-dividend ratios was due to variation in expected cashﬂows. Now it seems all price-dividend variation corresponds to discount-rate variation. We thought that the cross-section of expected returns came from the CAPM. Now we have a zoo of new factors.
In the beginning, there was chaos; practitioners thought one only needed to be clever to earn high returns. Then came the CAPM. Every clever strategy to deliver high average returns ended up delivering high market betas as well. Then anomalies erupted, and there was chaos again. The “value eﬀect” was the most prominent anomaly.
Fama and French (1993), (1996) brought order once again with size and value factors.
Explaining the value and size effect using value and size portfolios does work if you don't mind circular explanations. We have, as Fama stated in 1991, given a fishing license for factors, where every exception to the standard model is now a potential factor. It's a bit like the effects of Global Warming, except here you can say these are the one thousand things caused by variations in risk premiums.