Table 2 Subsample evaluation for real GDP growth: First release vs. f-curve. Root-mean-squared errors (RMSE) for forecasts on days with a new quarterly GDP release. A lower RMSE implies higher predictive accuracy. h=0 (h=1) denotes the forecast for the current (next) quarter. Panel (a) shows the evaluation for GDP figures that include the annual SFSO estimates (until 2018). Panel (b) excludes economic crises. As benchmark, we use the first quarterly release of the corresponding quarter. The Diebold-Mariano-West (DMW) test provides a p value for the null hypothesis of equal predictive accuracy against the alternative written in the column header (Diebold and Mariano 2002; West 1996). We assume a quadratic loss function
(a) Only when annual GDP estimates available (2000–2018) | ||||
|---|---|---|---|---|
RMSE | RMSE | Relative RMSE | DMW test (p value) | |
First release | f-curve | First release/f-curve | First release <f-curve | |
h=0 | 0.47 | 0.46 | 1.03 | 0.605 |
h=1 | 0.46 | 0.6 | 0.77 | 0.061 |
(b) Without crises (excluding 2008, 2009, 2020) | ||||
RMSE | RMSE | Relative RMSE | DMW test (p value) | |
First release | f-curve | First release/f-curve | First release <f-curve | |
h=0 | 0.4 | 0.4 | 1 | 0.504 |
h=1 | 0.4 | 0.43 | 0.91 | 0.166 |