Forecast | 1947–2005 | 2006–2017 |
|---|
Out-of-sample | \(E[d_{\cdot }]\) | \(Var[d_{\cdot }]\) | Diebold-Mariano | Out-of-sample |
|---|
\(R^2\) (%) | (\(\times 10^{-6}\)) | (\(\times 10^{-8}\)) | Test (p-value) | \(R^2\) (%) |
|---|
Robust (m) | 0.65 | 11.2 | 0.275 | 0.004 | 0.17 |
Proposed (a) | 0.64 | 11.1 | 0.519 | 0.040 | − 0.25 |
- This table reports forecasting performance measures for two sub-periods: 1947–2005 and 2006–2017. \(E[d_{i}]\) and \(Var[d_{i}]\) for \(i \in \{a, m\}\) denote the expected value and variance of loss difference, which represent average forecasting performance and its relative risk, respectively. Diebold-Mariano tests are to show if a forecast has significantly different predictability from a benchmark. All forecasts are free from look-ahead bias
