Generalized Cp Model Averaging for Heteroskedastic Models

This paper proposes a model averaging method, the generalized Mallows’ Cp (GC) method, which works well for heteroskedastic models. Under some regularity conditions, we provide a feasible form of the GC method and show that the GC method has asymptotic optimality not only as a model averaging method but also as a model selection method for heteroskedastic models. We perform some Monte Carlo studies to investigate the small sample properties of the GC method. The simulation results show that our method works well and performs better than alternative methods.