Forecasting with Bayesian multivariate vintage-based VARs

摘要: We consider the forecasting of macroeconomic variables that are subject to revisions, using Bayesian vintage-based vector autoregressions. The prior incorporates belief that, after first few data releases, subsequent ones likely consist revisions largely unpredictable. approach allows joint modelling more than one variable, while keeping concomitant increase in parameter estimation uncertainty manageable. Our model provides markedly accurate forecasts post-revision values inflation do other models literature.