Dynamic Panel Data

[ciru]

Hi,

I am estimating a dynamic panel data model using fixed effects (LSDV) as I have a fairly long time period. However, I suspect that there is cross section heteroskedasticity present in the data. If so, I understand from the EViews manual that in the Panel Options I should choose Cross Section Weights to correct for the presence of heteroskedasticity. This means that the regression will be estimated using Feasible GLS.

My question is, after choosing cross section weights to correct for the presence of heteroskedasticity is the regression model now known as a feasible GLS model or a fixed effects model?

Also, is feasible GLS appropriate for estimating dynamic panel data models?

Thanks.