Panel A: stock market returns ($$R_{oos}^{2}$$ in %) Predictor h=1 h=3 h=6 h=12 dp – 1.54 – 7.71 3.84 39.94* dy 0.26 – 3.80 11.39** 42.97* ep – 12.97 – 51.18 – 110.92 – 125.23 de – 23.58 – 117.28 – 281.21 – 221.84 svar 1.48 0.09 – 86.73 – 60.78 ntis – 4.51 – 17.89 – 48.71* – 91.04* ts – 1.90 – 5.84 – 6.79 5.11 ds – 10.95 – 82.37 – 313.15 237.55 Panel B: bond market returns ($$R_{\text {oos}}^{2}$$ in %) predictor h=1 h=3 h=6 h=12 dp – 2.23 – 9.52 – 6.33 – 12.03 dy – 2.01 – 6.55 – 4.47 – 16.25 ep – 3.00 – 6.68 – 11.86 – 26.72 de – 11.00 – 50.34 – 111.20 – 203.60 svar 2.12 – 15.41 – 27.75 – 4.92 ntis – 0.83 – 2.77 – 8.27 – 28.26** ts 4.20** 13.26** 25.37** 34.93** ds – 0.17 – 66.00 – 128.17 – 81.62
1. Notes: This table reports the out-of-sample R2 statistic ($$R_{\text {oos}}^{2}$$) proposed by (Campbell and Thompson 2008) from out-of-sample forecasts of Swiss stock market returns. This statistic obeys $$R_{oos}^{2}=1-\frac {\sum _{t=\text {tOOS}}^{T}(r_{t}-\hat {r}_{t})^{2}}{\sum _{t=\text {tOOS}}^{T}(r_{t}-\bar {r}_{t})^{2}}$$ in which $$\hat {r}$$ is the predicted value of the stock market excess returns and $$\bar {r}_{t}$$ is the historical mean of the return from the beginning of the sample until T−1. We test the statistical significance of $$R_{oos}^{2}$$ using the Clark and West (2007) test. A positive $$R_{\text {oos}}^{2}$$ indicates that the mean squared forecast error from the predictions by one of the forecast variables under study is lower than the benchmark, i.e., is lower than predictions using only the historical mean return. We evaluate the out-of-sample predictive ability of the predictors for the forecast period starting in January 2008 (tOOS), i.e., the evaluation period for the forecasts runs from January 1999 to December 2007. Then, we expand the window monthly from tOOS to T (December 2017). *, **, and *** indicate significance at the 10%, 5%, and 1% level, respectively. The predictor variables are described in Table 2 