EViews.com

Hello! I am working with eight North Eastern States of India over a period of 12 years 2000 to 2011. I am trying to estimate a two-way fixed effect model for my data set in panel setting, but when I do so, Eviews8 gives an error message.
My question is whether there is any requirement of the number of cross sections and number of periods that a two-way fixed effect model may be estimated, or where I may be wrong. Thanks in advance.
Vikas Dixit

What is the error message?

Thanks Glenn! Actually, I forgot to mention that when I am trying to estimate a two-way fixed effect model with cross section weights or cross section SUR (FGLS), the error message that I am getting is: "Period effect with cross section weights is not allowed". But when I leave the weights option unselected, that is, with no weights, I get the regression results. However, my feeling is that as it is the crossection - time series data, so I should take into account heteroskedasticity as well as contemporaneous correlation.

Hello! I am working with eight North Eastern States of India over a period of 12 years 2000 to 2011. I am trying to estimate a two-way fixed effect model for my data set in panel setting, but when I do so, Eviews8 gives an error message.
My question is whether there is any requirement of the number of cross sections and number of periods that a two-way fixed effect model may be estimated, or where I may be wrong. Thanks in advance.
Vikas Dixit

It's because the presence of the weighting means that the memory efficient algorithms for two-way fixed effects that we use cannot be used. You can estimate the model by including the period dummies explicitly. I will note that the number of observations that you have is quite small for the estimation of FGLS models.

Fine and many thanks Glenn! I will do it the way you told me.