Comparative Analysis of Fama and French Factor Models on the Russian Stock Market

In this paper we propose the verification of classical models of Fama and French as well as falsification of these models and comparison between them. The methodology used in this work makes it possible to carry out one of the most difficult tasks, namely, to evaluate models on the Russian stock market in the realities of the modern financial world. Using different sorting methods, change of the percentiles, chaotic construction of factors and replacing the traditional BV/MV multiplier, different models are tested and their comparative analysis is carried out. The models are compared in accordance with the coefficients of determination and GRS statistics. We find a size effect on the Russian stock market, but the value effect is very unstable when changing sorting and percentiles, therefore, another factor that correctly describes yields is needed. The model in which portfolios are formed by the P/E works better than the classical model of Fama and French. The testing of the five-factor model has shown better results compared to the three-factor model, and better applicability in portfolios formed by the indicators of book value and investment. The results obtained from the testing of three-factor models with the inclusion of portfolios built on various indicators - BV/MV, investments and profitability – has shown the high applicability of models with the inclusion of the investment indicator. It means that the investment indicator can be used as a reliable tool for analysis, since it reflects the current position of the company and its prospects. Further research should reveal the representativeness of the results and the applicability of the models on other capital markets similar to the Russian market.