Trade offs between efficiency and robustness in the empirical evaluation of asset pricing models

Even though we can formally express an asset pricing model in a classical Beta or in the relatively new stochastic discount factor representation, their key empirical features – notably, efficiency and robustness – may differ when estimated by the generalized method of moments. In particular, we find that there appears to be a degree of trade-off between these features, which are linked to the asset pricing representation. Here, we investigate this issue on single and multi-factor asset pricing models using US and UK data. In this setting, our results show that the SDF approach is more likely to be less efficient but more robust than Beta method. We show that the main drivers of this trade-off are the higher order moments of the factors, in which skewness plays an important role, and especially the covariance between returns and factors.