Hülsewig, Oliver Mayr, Johannes Wollmershäuser, Timo
Year of Publication:
CESifo working paper 2371
This paper proposes a new method of forecasting euro area quarterly real GDP that uses area-wide indicators, which are derived by optimally pooling the information contained in national indicator series. Following the ideas of predictive modeling, we construct the area-wide indicators by utilizing weights that minimize the variance of the out-of-sample forecast errors of the area-wide target variable. In an out-of-sample forecast experiment we find that our optimal pooling of information approach outperforms alternative forecasting methods in terms of forecast accuracy.
Forecasting aggregation model averaging real time experiment