A0219
Title: Testing model-based contributions in misspecified asset pricing models
Authors: Cisil Sarisoy - Federal Reserve Board (United States) [presenting]
Abstract: Test statistics designed to evaluate the risk-premium contributions of asset pricing models are developed for individual assets. These statistics are robust to model misspecification, allowing for reliable use in both correctly and misspecified models. The method is applied to the well-known asset pricing models that include both traded and non-traded factors. The empirical results show that, in testing model-based contributions of asset pricing models, the misspecification-robust test statistics developed can lead to economically different conclusions from those derived under correctly specified models.
