A0547
Title: Comparing factor models with conditioning information
Authors: Seok Young Hong - Lancaster University Management School (United Kingdom) [presenting]
Abstract: A novel framework is developed to conduct asymptotically valid tests for comparing factor models with conditioning information. The tests are based on a metric analogous to the squared Sharpe ratio improvement measure that is used to gauge the extent of model mispricing in an unconditional setting. An estimator for the metric is proposed, and its limiting properties, establishing the asymptotic normality, are studied. An advantage of our framework is that it can be applied without an a priori knowledge of the persistent nature of the conditioning variables. A range of dependence classes is accommodated, including stationary, nearly stationary, integrated, and local-to-unity.
