| Â |
GARCH model
|
EGARCH model
|
|---|
| Â | Pre-crises | Post-crises | Â | Pre-crises | Post-crises |
|---|
α0
| 0.154969** (0.026322) | 0.114691** (0.017558) | α0
| 0.153675** (0.025096) | 0.089605** (0.019572) |
α1
| 0.068317** (0.025540) | 0.140572** (0.024231) | α1
| 0.061981* (0.02573) | 0.147118** (0.023669) |
ψ
| 0.014185 (0.010686) | 0.018941** (0.005445) |
ψ
| 0.021001* (0.010061) | 0.035986** (0.003204) |
β0
| 0.197274** (0.017770) | 0.029230** (0.003107) | β0
| −0.23126** (0.019195) | −0.231288** (0.017791) |
β1
| 0.236202** (0.021655) | 0.181986** (0.014402) | β1
| 0.398443** (0.029732) | 0.299825** (0.025086) |
| Â | Â | Â |
φ
| −0.074287** (0.017424) | −0.162797** (0.016946) |
β2
| 0.674075** (0.021069) | 0.801902** (0.010664) | β2
| 0.861252** (0.011513) | 0.939410** (0.006352) |
δ
(OP)
| −0.049991** (0.011195) | 0.008880** (0.000742) |
δ
(OP)
| −0.024189** (0.009345) | 0.001229 (0.005441) |
ARCH LM | 0.139988 [0.7082] | 0.242116 [0.6267] | Â | 0.422741 [0.5157] | 0.032083 [0.8579] |
- The figures in parenthesis are standard errors. The figures in braces are probabilities. The ARCH LM test is applied on residuals of each model for exploring autocorrelation and heteroskedasticity problem after estimation of models
- * and ** indicates p < 5% and 1% respectively
- The GARCH model is given as: \( {\mathrm{R}}_{\mathrm{t}\;\left(\mathrm{PSE}\right)}={\upalpha}_0+{\upalpha}_1\;{\mathrm{R}}_{\mathrm{t}\hbox{-} 1}+\psi {\mathrm{R}}_{\mathrm{t}\hbox{-} 1\;\left(\mathrm{OP}\right)}+{\upvarepsilon}_{\mathrm{t}},\kern0.6em {\mathrm{h}}_{\mathrm{t}\;\left(\mathrm{PSE}\right)}={\upbeta}_0+{\upbeta}_1\;{\upvarepsilon}_{\mathrm{t}\hbox{-} 1}^2+{\upbeta}_2\;{\mathrm{h}}_{\mathrm{t}\hbox{-} 1}+{\delta}_{\left(\mathrm{OP}\right)} \)
- The EGARCH model is given as: \( {\mathrm{R}}_{\mathrm{t}\;\left(\mathrm{PSE}\right)}={\upalpha}_0+{\upalpha}_1\;{\mathrm{R}}_{\mathrm{t}\hbox{-} 1}+\psi\;\mathrm{O}{\mathrm{P}}_{\mathrm{t}\hbox{-} 1}+{\upvarepsilon}_{\mathrm{t}},\kern0.6em \ln {\mathrm{h}}_{\mathrm{t}\;\left(\mathrm{PSE}\right)}={\upbeta}_0++{\upbeta}_1\;\left|\frac{\in_{\mathrm{t}}\hbox{-} 1}{\sqrt{{\mathrm{h}}_{\mathrm{t}\hbox{-} 1}}}\right|+\varphi\;\frac{\in_{\mathrm{t}}\hbox{-} 1}{\sqrt{{\mathrm{h}}_{\mathrm{t}\hbox{-} 1}}}+{\upbeta}_2\;{\mathrm{h}}_{\mathrm{t}\hbox{-} 1}+{\delta}_{\left(\mathrm{OP}\right)} \)
