Forecasting with High Dimensional Panel VARs - Prof. Gary Koop

Abstract:

In this paper, we develop econometric methods for estimating large Bayesian time-varying parameter panel vector autoregressions (TVP-PVARs) and use these methods to forecast inflation for euro area countries.  Large TVP-PVARs contain huge numbers of parameters which can lead to over-parameterization and computational concerns.  To overcome these concerns, we use hierarchical priors which reduce the dimension of the parameter vector and allow for dynamic model averaging or selection over TVP-PVARs of different dimensions and different priors.  We use forgetting factor methods which greatly reduce the computational burden.  Our empirical application shows substantial forecast improvements over plausible alternatives.