Table 3.

In-sample estimation results.

	Model 1	Model 2	Model 3	Model 4
Panel A: h = 1
${\beta }_0/{\beta }_{0,S_0}$	−1.721∗∗∗	−1.686	−1.997∗∗∗	−1.743∗∗∗
${\beta }_{0,S_1}$			−1.200∗∗∗	−2.445∗∗∗
${\beta }_1/{\beta }_{1,S_0}$	0.797∗∗∗	0.675	0.765∗∗∗	0.523∗∗∗
${\beta }_{1,S_1}$			0.855∗∗∗	0.461∗∗∗
${\gamma }_1/{\gamma }_{1,S_0}$		0.141		0.293∗∗
${\gamma }_{1,S_1}$				0.321∗
$P^{00}$			0.996	0.996
$P^{11}$			1.000	1.000
Panel B: h = 5
${\beta }_0/{\beta }_{0,S_0}$	−3.060∗∗∗	−2.065∗∗∗	−4.754∗∗∗	−5.947∗∗∗
${\beta }_{0,S_1}$			−2.915∗∗∗	−3.571∗∗∗
${\beta }_1/{\beta }_{1,S_0}$	0.629∗∗∗	0.487∗∗∗	0.367∗∗∗	0.443∗∗∗
${\beta }_{1,S_1}$			0.678∗∗∗	0.594∗∗∗
${\gamma }_1/{\gamma }_{1,S_0}$		0.359∗		0.932∗∗∗
${\gamma }_{1,S_1}$				0.019
$P^{00}$			0.923	0.940
$P^{11}$			0.955	0.961
Panel C: h = 20
${\beta }_0/{\beta }_{0,S_0}$	−6.611∗∗∗	−7.930∗∗∗	−9.807∗∗∗	−13.193∗∗∗
${\beta }_{0,S_1}$			−5.442∗∗∗	−2.431∗∗∗
${\beta }_1/{\beta }_{1,S_0}$	0.203∗∗∗	0.212∗∗∗	0.221∗∗∗	0.954∗∗∗
${\beta }_{1,S_1}$			0.384∗∗∗	0.623∗∗∗
${\gamma }_1/{\gamma }_{1,S_0}$		−0.295∗∗		0.660∗∗∗
${\gamma }_{1,S_1}$				0.163
$P^{00}$			0.975	0.971
$P^{11}$			0.974	0.972

Notes: This table shows the parameter estimation results from an in-sample perspective. ${\beta }_0$is the constant term. ${\beta }_1$and ${\gamma }_1$present the coefficient of RV and OVX respectively. $S_0$and $S_1$indicate the low and high volatility respectively. $P^{00}$and $P^{11}$are the probability value of the transition matrix. Different horizons h = 1, 5, 20 are shown in Panel A, B, C as well. Asterisk ∗∗∗, ∗∗ and ∗ denote rejections of null hypothesis at 1%, 5% and 10% level. In this table, p value less than 0.1 are meaningful.