Comparing TGARCH parameter estimates

[yonaaz]

Hi guys,

I’m struggling with finding a way to compare my parameter estimates when I apply an asymmetric GARCH (T-GARCH). I have a serie of stock returns and have established that there are volatility asymmetries present. I have also composed several stock portfolios based on company specifics. When I apply the TGARCH to the different portfolios, some display significant asymmetries and some don’t. Is there any formal test I can apply to be able to conclude that the asymmetry is dependent on company specifics?

As of now I’m using a univariate model, do I need to use a multivariate model instead?

I have also added an additional exogenous variable to the conditional variance equation which is significant in some of the portfolios and insignificant in other. I would like to be able to compare that estimate between portfolios as well.

I’m using eviews 7

Very greatful for any help!



Best regards
Yonaaz

[Triston]

I think you can include a dummy variable for TGARCH in the variance regressor text box. Let me know if that worked for you. Cheers!