Skip to main content

Table 5 In-sample predictability of Swiss stock market excess returns: the US predictors versus GVD

From: What Goliaths and Davids among Swiss firms tell us about expected returns on Swiss asset markets

Panel A: Predictive power of US predictors
Predictor\(\hat {\beta }\)R2(%)\(\hat {\beta }\)R2(%)\(\hat {\beta }\)R2(%)\(\hat {\beta }\)R2(%)
(p value)(0.23) (0.19) (0.11) (0.06) 
(p value)(0.12) (0.12) (0.06) (0.05) 
(p value)(0.32) (0.42) (0.45) (0.38) 
(p value)(0.52) (0.43) (0.35) (0.36) 
svar– 0.67***2.83– 0.66***6.47– 0.312.380.010.00
(p value)(0.00) (0.00) (0.25) (0.45) 
(p value)(0.35) (0.32) (0.36) (0.37) 
ts– 0.270.48– 0.180.47–
(p value)(0.83) (0.70) (0.55) (0.25) 
ds– 0.411.06– 0.230.80–
(p value)(0.78) (0.60) (0.48) (0.32) 
Panel B: marginal contribution of GVDnew in the presence of best US predictor at each forecast horizon
 \(\hat {\beta }\)Partial R2(%)\(\hat {\beta }\)Partial R2(%)\(\hat {\beta }\)Partial R2(%)\(\hat {\beta }\)Partial R2(%)
(p value)(0.07) (0.04) (0.04) (0.10) 
  1. Notes: Panel A of this table presents OLS estimates from univariate regressions of h-month ahead Swiss stock market returns on each potential US predictor variable described in Tables 1 and 2 with the exception of US GVD. Panel B gives OLS estimates from regressions of h-month ahead Swiss stock market returns on GVDnew and the best US predictor at each forecast horizon. These estimates show whether GVDnew adds predictive power in the presence of the best US predictor. The sample period runs from January 1999 to December 2017. We compute heteroskedasticity and autocorrelation robustt-statistics (in parentheses below the estimates) from a wild bootstrap procedure that tests the null hypothesis of \(\hat {\beta }^{h}=0\) against the alternative that \(\hat {\beta }^{h}>0\) or \(\hat {\beta }^{h}<0\) in the case of svar. *, **, and *** indicate significance at the 10%, 5%, and 1% level, respectively