Robust Inference in Linear Asset Pricing Models

Many asset pricing models include risk factors that are only weakly correlated with the asset returns. We show that in the presence of a factor that is independent of the returns ("useless factor"), the standard inference procedures for evaluating its pricing ability could be highly misleading in misspecified models. Our proposed model selection procedure, which is robust to useless factors and potential model misspecification, restores the standard inference and proves to be effective in eliminating factors that do not improve the model's pricing ability. The practical relevance of our analysis is illustrated using simulations and an empirical application