You have to mention panel correlation and autocorrelation option in xtgls function at the end, the model you have run at the end is wrong because it is assuming no heteroskedesticity and no autocorrelation is present.
Thank you very much for this useful tutorial. I have a question please. Why did you take some variables in the first difference when you have run the FGLS MODEL
Respected Sir, I am working on panel data and for this I have two equations. I equation 1 I first checked cross sectional independence and then applied the unit root tests regarding to it. But, in equation 2 I have no cross sectional independence and I not able to find commands for the unit root tests (except pesaran test) for this type of equation. Kindly, guide me in this Topic. Thank you in anticipation.
yes, I did and found that my dependent variable roa and two independent variables had unit root so I took 1st difference of them. it can be seen in the video when I run fgls ( d.roa, d.sdlt and d.etl).
@@everythingnyimbo This can be done by the Hausman test, but there are other tests, for example in ivreg2 you can use the orthog/endog options, to test for the exogeneity of one or multiple variable.
You have to mention panel correlation and autocorrelation option in xtgls function at the end, the model you have run at the end is wrong because it is assuming no heteroskedesticity and no autocorrelation is present.
Can I use FGLS with Fixed effects at the same time?
Thank you very much for this useful tutorial. I have a question please. Why did you take some variables in the first difference when you have run the FGLS MODEL
to check for autocorrelation.
Respected Sir, I am working on panel data and for this I have two equations. I equation 1 I first checked cross sectional independence and then applied the unit root tests regarding to it. But, in equation 2 I have no cross sectional independence and I not able to find commands for the unit root tests (except pesaran test) for this type of equation. Kindly, guide me in this Topic. Thank you in anticipation.
I read Wooldridge (2002) and He told about FE GLS. Is there any command in STATA to perform FE GLS?
I have the exact same question! Do you find have solutions? Thank you!
Did you check for unit root before running xtgls regression?
yes, I did and found that my dependent variable roa and two independent variables had unit root so I took 1st difference of them. it can be seen in the video when I run fgls ( d.roa, d.sdlt and d.etl).
MadeEasy yes i just wanted to be sure :) after fgls are there further tests?
Yes, panel data unit root tests. This is why, the variables are labelled as d.roa, d.sdlt etc.
can i use FGLS in eviews??
Yes, eviews10. I dont know about the earlier versions of Eviews
how about endogeneity?
endogeneity of cross sections must be tested.
@@madeeasy7691 whats the command for this?
@@everythingnyimbo This can be done by the Hausman test, but there are other tests, for example in ivreg2 you can use the orthog/endog options, to test for the exogeneity of one or multiple variable.