Forecasting GDP growth in times of crisis: private sector forecasts versus statistical models

This paper examines the accuracy of short run forecasts of Dutch GDP growth by several linear statistical models and private sector analysts. We focus on the financial crisis of 2008-2009 and the dot-com recession of 2001-2002. The dynamic factor model turns out to be the best model. Its forecast accuracy during the crisis deteriorates much less than that of the other linear models and hardly at all when backcasting and nowcasting. Moreover, the dynamic factor model beats the private sector forecasters at nowcasting. This finding suggests that adding judgement to a mechanical model may not improve short-term forecasting performance.

Keywords: Nowcasting, Professional Forecasters, Factor Model, Forecasting.
JEL Classifications: E52, C53, C33.

Working paper no. 320