Skip to main content

Table 8 In-Sample R-squares Reported by VAR(q) and ARCH-in-Mean over Different Time Horizons

From: Timing the market: the economic value of price extremes

R in 2 Panel A. Monthly Data Observations Panel A. Quarterly Data Observations
1950.01-1985.12 1986.01-2015.12 1950.01-2015.12 1950.01-1985.12 1986.01-2015.12 1950.01-2015.12
VAR(q) (%) 1.89 1.17 1.57 12.54 2.60 6.76
ARCH-in-Mean (%) 0.32 0.27 0.16 1.37 0.42 0.82
Predictability Ratio 5.91 4.33 9.81 9.15 6.19 8.20
  1. Note. This table reports the in-sample R-squares, Rin2 reported by VAR(q) and ARCH-in-Mean. In-sample predictability is performed over different time horizons for the sake of being robust. Since ARCH tests report no heteroscedasticity in quarterly returns, thus we assume constant volatility for quarterly returns. Predictability ratio is calculated as the ratio of in-sample R-square of VAR(q) over the in-sample R-square of ARCH-in-Mean