Combining Bayesian VARs with Survey Density Forecasts: Does it Pay Off?

This paper studies how to combine real-time forecasts from a broad range of Bayesian vector autoregression (BVAR) specifications and survey by optimally exploiting their properties. To do that, it compares the forecasting performance optimal pooling tilting techniques, including for predicting euro area inflation GDP growth at medium-term forecast horizons using both univariate multivariate metrics. Results show that Survey Professional Forecasters (SPF) provides good point performance, but also SPF perform poorly in terms densities all variables horizons. Accordingly, when model combination or individual models are tilted SPF's first moments, accuracy calibration improve, whereas they worsen second moments included. We conclude judgement incorporated can considerably increase accuracy, however, way extent which is matters.