Page 1 of 1

Overflow when making residuals series

Posted: Sat Oct 21, 2017 6:18 am
by a.do88
Hi

I got an "overflow" message when making a residuals series from a bivariate E-GARCH-M with student's t distribution error for stock returns (first difference log). It does not happen to all pairs of countries in my sample, but only to some. In addition, when I play around with the model and change the specifications (e.g. use a general EGARCH with normal distribution errors - the issue disappears). I have read some posts about overflow errors in Eviews forum, but just want to double check is it because the estimated residuals from the initial model are getting too big? Does it mean the first model is not a good model for some countries?

Thanks in advance for your help.

Regards, Anh

Re: Overflow when making residuals series

Posted: Wed Oct 25, 2017 8:32 am
by a.do88
Can anyone please help me :(

Re: Overflow when making residuals series

Posted: Wed Oct 25, 2017 12:20 pm
by EViews Glenn
Are these residuals post-estimation or during estimation?

Re: Overflow when making residuals series

Posted: Thu Oct 26, 2017 5:01 am
by a.do88
they are post-estimations I guess...so I run the EGARCH model for two countries, then I make the residuals from the model..

Re: Overflow when making residuals series

Posted: Mon Oct 30, 2017 10:28 am
by EViews Glenn
We're going to have to see the workfile with the estimated system object.