Improving Model-Based Near-Term GDP Forecasts by Subjective Forecasts: A Real-Time Exercise for the G7 Countries

Abstract

We investigate to what extent it is feasible to improve model-based near-term GDP forecasts by combining them with judgmental (quarterly) forecasts by professional analysts (Consensus survey) in a real-time setting. Our analysis covers the G7 countries over the years 1999-2013. We consider as combination schemes the weighted average and the linear combination. Incorporating subjective information delivers sizable gains in forecasting ability of statistical models for all countries except Japan in 1999-2013, even when subjective forecasts are somewhat dated. Accuracy gains are much more pronounced in the volatile period after 2008 due to a marked improvement in predictive power of Consensus forecasts. Since 2008, Consensus forecasts are superior at the moment of publication for most countries. For some countries Consensus forecasts can be enhanced by model-based forecasts in between the quarterly release dates of the Consensus survey, as the latter embody more recent monthly information.