GLS and fixed effects

#1
Hi,

I'm pretty new to Stata so am having a few problems. I am working with panel data and would like to run a regression of the following form:

xst = αs + βt + λyst + δlst-4 + εst

where x is a measure of poverty in state s at time t, αs is a state fixed effect, βt is a year dummy variable, yst is a vector of variables trated as exogenous, and lst-4 is a measure of land reforms lagged four periods. I would like to model εst as an AR(1) process (where the degree of autocorrelation is state specific) with heteroscedasticity between states (assigning analytical weights according to the estimated variance of each states residuals).

I am having problems since the options on Stata to fit panel data using GLS don't give me an option to control for fixed effects. I'd really appreciate it if someone could tell me how to estimate the above equation via GLS whilst also controlling for fixed and time effects, or if this isn't directly possible another method that would achieve similar results.

Thanks,

Lily
 
#2
Hey there,

Have you tried the "xtregar" command? I ask because I know that xtgls would not fit the model you want to estimate but it seems that xtregar does.

Etienne