Table 4.
Pseudo-out-of-sample forecast evaluation.
| (a) Real GDP growth | ||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| RMSFE |
ACRPS |
ALPL |
||||||||||
| BVAR | 0.52 | 0.95 | 1.37 | 1.75 | 0.26 | 0.48 | 0.71 | 0.93 | −0.83 | −1.38 | −1.77 | −2.07 |
| BVAR | 0.51 | 0.94 | 1.36 | 1.74 | 0.25 | 0.46 | 0.69 | 0.90 | −0.74 | −1.24 | −1.64 | −1.92 |
| (I) Pre-COVID-19 period: 1995:Q1–2019:Q4 | ||||||||||||
| BVAR | 9.03 | 11.72 | 13.07 | 14.79 | 5.73 | 7.14 | 8.22 | 9.26 | −34.09 | −63.40 | −55.06 | −44.43 |
| BVAR | 7.38 | 8.71 | 8.90 | 9.69 | 5.01 | 5.61 | 6.13 | 7.35 | −6.67 | −6.72 | −7.43 | −8.89 |
| BVAR | 7.49 | 8.93 | 9.04 | 9.77 | 5.05 | 5.76 | 6.08 | 7.04 | −61.45 | −80.80 | −69.11 | −63.25 |
| (II) COVID-19 period: 2020:Q1–2021:Q2 | ||||||||||||
| (b) HICP inflation | ||||||||||||
| RMSFE |
ACRPS |
ALPL |
||||||||||
| BVAR | 0.30 | 0.54 | 0.75 | 0.99 | 0.16 | 0.28 | 0.39 | 0.52 | −0.25 | −0.82 | −1.12 | −1.39 |
| BVAR | 0.30 | 0.53 | 0.72 | 0.96 | 0.16 | 0.28 | 0.38 | 0.50 | −0.22 | −0.72 | −1.03 | −1.33 |
| (I) Pre-COVID-19 period: 1995:Q1–2019:Q4 | ||||||||||||
| BVAR | 0.61 | 1.06 | 1.35 | 1.56 | 0.32 | 0.66 | 0.78 | 0.80 | −1.09 | −1.90 | −1.68 | −1.59 |
| BVAR | 0.58 | 0.99 | 1.12 | 1.14 | 0.31 | 0.64 | 0.71 | 0.67 | −0.86 | −1.71 | −1.67 | −1.58 |
| BVAR | 0.59 | 1.02 | 1.16 | 1.21 | 0.31 | 0.63 | 0.68 | 0.67 | −1.05 | −1.82 | −1.55 | −1.48 |
| (II) COVID-19 period: 2020:Q1–2021:Q2 | ||||||||||||
Note: Forecast evaluation metrics prior to and during the COVID-19 pandemic for (a) real GDP growth and (b) HICP inflation. BVAR denotes a standard Gaussian BVAR, BVAR is the fat-tailed BVAR, and BVAR is the BVAR augmented with dummy variables for each quarter in the COVID-19 period.