This thesis explains a methodology for relative tests of factor-based asset pricing models. In particular, the concept of comparing competing models with the maximum Sharpe ratio is illustrated, both for nested and non-nested models. For empirical purposes, the thesis explains the use of spanning regressions, the GRS test, bootstrapping and the asymptotic behaviour of Sharpe ratio differences. Applying the methodology to data samples of the US and the Japanese market for popular models, the Stambaugh/Yuan four-factor model shows the best performance for the US while a two-factor model consisting of the value factor HML and the investment factor RMW by Fama/French is to be preferred for Japan.      «

This thesis explains a methodology for relative tests of factor-based asset pricing models. In particular, the concept of comparing competing models with the maximum Sharpe ratio is illustrated, both for nested and non-nested models. For empirical purposes, the thesis explains the use of spanning regressions, the GRS test, bootstrapping and the asymptotic behaviour of Sharpe ratio differences. Applying the methodology to data samples of the US and the Japanese market for popular models, the S...     »