Basic ARCH/GARCH question
Posted: Sat Oct 12, 2013 8:36 am
Hi all,
I've used ARCH modeling before in Matlab and G@RCH, but am a complete newbie to it in eviews. I have a dataset that is basically a highly unbalanced dated panel, with monthly data for 25 countries with different starting dates for each country's data series (i.e. financial risk indicators are available for one country from 1998, for another from 1992, for another 2005, etc.). I've used eviews quite a lot for panel analysis, but never for ARCH modeling. My questions, insanely basic though they are
, would benefit from some help:
1. How does one structure a workfile/database to utilize the ARCH modeling in the equation estimation dialogue? When I import as a dated panel, it doesn't allow it (and I saw in another post that ARCH isn't available with panels).
2. After importing the data, is there a way to make sure that it treats the country as an identifier? That is, I don't want to have eviews treat the 2847 observations I have as one long stream, I want it to differentiate that these are independent streams?
3. Does the fact that I have different starting times for the key financial risk variable cause a problem?
Thanks so much, sorry that these questions are basic, but I've figured them out in other programs but not eviews!
I've used ARCH modeling before in Matlab and G@RCH, but am a complete newbie to it in eviews. I have a dataset that is basically a highly unbalanced dated panel, with monthly data for 25 countries with different starting dates for each country's data series (i.e. financial risk indicators are available for one country from 1998, for another from 1992, for another 2005, etc.). I've used eviews quite a lot for panel analysis, but never for ARCH modeling. My questions, insanely basic though they are
, would benefit from some help:1. How does one structure a workfile/database to utilize the ARCH modeling in the equation estimation dialogue? When I import as a dated panel, it doesn't allow it (and I saw in another post that ARCH isn't available with panels).
2. After importing the data, is there a way to make sure that it treats the country as an identifier? That is, I don't want to have eviews treat the 2847 observations I have as one long stream, I want it to differentiate that these are independent streams?
3. Does the fact that I have different starting times for the key financial risk variable cause a problem?
Thanks so much, sorry that these questions are basic, but I've figured them out in other programs but not eviews!