Table 2.11: Significance after correction for dependence: an estimation based approach.
Panel A: This table shows for all sets of trading rules applied to the LIFFE cocoa series in the period 1983:1-1987:12 the percentage of trading rules for which the estimated coefficient of the buy (sell) dummy in the regression function of the exponential garch model is significantly positive (negative) at the 10% significance level with a one sided test (second and third column). The fourth column shows the percentage of trading rules for which the coefficient of the buy dummy is significantly positive and the coefficient of the sell dummy is significantly negative.
Rule tBuy>tc tSell<-tc tBuy>tc ∧ tSell<-tc MA 40.2 32.8 29.6 TRB 41.9 22.7 16.6 Filter 38.7 17.5 9.8 All 40.6 27.4 22.8
Panel B: This table shows for all sets of trading rules applied to the LIFFE cocoa series in the period 1983:1-1987:12 the percentage of trading rules for which the estimated coefficient of the buy (sell) dummy in the regression function of the exponential garch model is significantly negative (positive) at the 10% significance level with a one sided test (second and third column). The fourth column shows the percentage of trading rules for which the coefficient of the buy dummy is significantly negative and the coefficient of the sell dummy is significantly positive.
Rule tBuy<-tc tSell>tc tBuy<-tc ∧ tSell>tc MA 3.6 4.1 1.5 TRB 5.2 9.6 1.9 Filter 2.1 6.8 0.7 All 4.0 6.4 1.6
84