Specifying a Bayesian vector autoregression for short-run macroeconomic forecasting with an application to Finland

Abstract

The aim of this paper is to specify a small econometric model capable of generating adjustment-free, short-run forecasts of key macroeconomic variables on a monthly basis. The aim is carried out using the vector autoregression approach in conjunction with a Bayesian specification procedure. The Bayesian approach to forecasting is reviewed and applied using Finnish data from the 1980s. The out-of-sample forecasting performance of the model is found to be satisfactory.