Table 3. Regime-switching modelling of implied volatility | ||||
Model 1 |
Model 2 |
Model 3 |
Model 4 | |
Model parameters |
I | |||
w1 |
0.3106a |
0.1522a |
0.3030a |
0.0048a |
w 2 |
0.2080a |
0.0088a |
0.2057a |
0.3494a |
δι |
0.7995a |
0.9747a | ||
δ2 |
0.9611a |
-0.0693a | ||
βι |
-0.0339 |
0.5529a |
-0.1489a |
-0.3231a |
β2 |
0.0044 |
0.0072 |
-0.0991 |
0.8789a |
Y1 |
20.5833a |
5.7630a | ||
γ 2 |
11.6824a |
-24.6102a | ||
Ψ1 | ||||
Ψ2 | ||||
Hypothesis tests |
I | |||
w1 = w 2 |
4590.3206a |
1764.6474a |
3844.3931a |
10957.0160a |
δι = δ2 |
225.7556a |
9901.6532a | ||
βι = β2 |
0.1639 |
85.26211a |
0.3771 |
430.3168a |
Yi = γ 2 |
28.2025a |
432.8464a | ||
Φ1 = Ψ2 |
4.5052b |
5.4429b | ||
P11 = P 22 |
20.8045a |
175.5199a | ||
LL |
6822.32 |
11007.45 |
6883.55 |
11386.28 |
Notes: Significance at the 1, 5 and 10 % level is denoted by a, b and |
respectively. The estimated Markov regime-switching models
are represented by vt = wi + βirt-1 +ζt for Model 1,
vt = wi + δivt-1 + βirt-1 +ζt for Model 2,
vt = wi + βirt-1 + γi rt2-1 +ζt for Model 3, and
vt = wi +δivt-1 + βirt-1 + γi rt2-1 +ζt for Model 4. The null
hypothesis tests are distributed asχ2(1) . LL is the log maximum
likelihood function.
23