Article ID: | iaor20013981 |
Country: | Netherlands |
Volume: | 17 |
Issue: | 2 |
Start Page Number: | 181 |
End Page Number: | 201 |
Publication Date: | Apr 2001 |
Journal: | International Journal of Forecasting |
Authors: | Dpke Jrg |
Keywords: | time series & forecasting methods |
The paper elaborates on the sources of macroeconomic forecast errors in Germany covering a time period ranging from 1963 to 1999. The joint predictions of the so-called ‘six leading’ research institutes are analyzed. The forecast errors are discussed within an aggregate demand/supply scheme. Structural vector autoregressive models are estimated to identify the shocks underlying the business cycle. The empirical results suggest that, in general, the shocks are helpful in explaining the forecast errors. However, the correlations are rather weak. In addition, lagged shocks help also to explain the forecast errors of the institutes. This is in line with previous evidence that forecasters' expectations are not rational.