Noisy prices and the Fama-French five-factor asset pricing model in China

This paper provides an empirical evaluation of the five-factor model recently presented by Fama and French (2015a) that augments the traditional three-factor model with two new factors related to profitability and investment, taking into account the bias in mean returns induced by noise in prices. Using an extensive sample over the period 1997 to 2015, we find that the five-factor model consistently outperforms the three-factor model in the Chinese equity market. In contrast to the findings in Fama and French (2015a), both value and profitability factors are important, while the investment factor is found to be redundant for describing average returns in our sample. With respect to the double-sorted left-hand-side (LHS) portfolios, the main problem with the five-factor model is its failure to fully capture the high average returns of stocks whose returns perform like those of growth firms that invest conservatively due to low profitability. On the other hand, although we find mixed results in the three-dimensional sorting, the LHS portfolios with extremely low average returns are those that cause serious problems for the five-factor model.