Page 1 of 1
OLS on Panel data is random effect?
Posted: Thu Sep 27, 2012 1:30 am
by giotira
Hi to all,
a really technical question. I have an unbalanced panel data set, defined as panel data on Eviews.
When I am running quick\equation\LS am I doing a random effect estimation on the data set? if not what am I doing then? moreover how can I do a random effect estimation?
many thanks
Re: OLS on Panel data is random effect?
Posted: Thu Sep 27, 2012 7:44 am
by EViews Gareth
There is an option on the Panel Options part of the estimate dialog that lets you specify random effects.
Re: OLS on Panel data is random effect?
Posted: Thu Sep 27, 2012 8:04 am
by giotira
So If I am not specyfing it, Eviews is simply using an OLS estimation on the whole sample, that is it does not consider the data as a panel date and every observation is a different event. is this right?
Re: OLS on Panel data is random effect?
Posted: Thu Sep 27, 2012 8:06 am
by EViews Gareth
Essentially.
Re: OLS on Panel data is random effect?
Posted: Thu Sep 27, 2012 8:13 am
by giotira
What is then the difference with a GLS?
simply the autoregressive component on the residuals?
Re: OLS on Panel data is random effect?
Posted: Thu Sep 27, 2012 11:42 am
by EViews Glenn
Not sure I understand the question. The OLS on the stacked panel is just that. The only thing about the panel that is relevant is lag handling.