Problems with some GARCH I am trying to run

[HETISFRANK]

Hello,

For my thesis I was given the assignment to research the effect of macroeconomic news on stocks. I chose to investigate the effect of US announcements coming from the BLS on the Brazilian stock market. In my case I wanted to check if the announcements would have a significant effect on the volatilities of the indices.

From previous researches I concluded that the best way of achieving this was using a GARCH time-series to get my volatilities and then test significance with an F-test. Unfortunately I am unfamiliar with this technique. Both the actual reasoning behind it and applying it with the help of statistical software packages.

The data I currently have at my disposal are both the price and returns of several Brazilian indices, dummy variables for the most important US macroeconomic news announcements and in most cases the change percentage of those same announcements.

I've found a paper by Nikkinen et al. (2006) that investigates the same effect just with a different sample and I've been trying to use this paper to get my methodology. So far it has cleared a few things up for me and I have a vague idea what they are doing but at the same time a lot of things are still unclear to me. So any help on both the reasoning and/or how I'd get the correct time-series from my available software package similar to the above mentioned methodology, in Stata or EViews, would be much appreciated. Personally I am more experienced with Stata although by no means an expert at it. When confronting my supervisor with my problems she suggested I try EViews and I've tried it on campus PCs but so far it didn't really make any sense to me.

Thanks in advance, Frank

P.S: I've uploaded the pages which describe the methodology in the Nikkinen et al. (2006) paper. I wanted to upload the entire thing but I don't want to get any weird copyright cases going against me. If things are still unclear, just tell me and I'll try to provide more information.

Nikkinen et al. (2006) - Page 1

Nikkinen et al 1.jpg (454.82 KiB) Viewed 2181 times

Nikkinen et al. (2006) - Page 2

Nikkinen et al 2.jpg (506.94 KiB) Viewed 2181 times

Nikkinen et al. (2006) - Page 3

Nikkinen et al 3.jpg (410.68 KiB) Viewed 2181 times

[trubador]

You should refer to textbooks and/or online educational sources to understand the theory behind the GARCH-type modeling. As for the EViews implementation, your job is fairly easy. All you have to do is consider your explanatory variables as variance regressors. Following example might give you the idea: http://forums.eviews.com/viewtopic.php?f=6&t=7677

[HETISFRANK]

Thank you for the fast response. I know textbooks are my best bet to understand the theory behind the modeling. Thing is that I'll get proper classes on the subject next year if I go into my finance Master. So far it hasn't been touched upon in my Bachelor program but still I got stuck with this assignment for my thesis.

Anyways, I'll check out the link and try and go from there to see if I actually get the volatilities I need.