Hello,
After Hausman test when I run fixed effects model, I have autocorrelation, heteroscedasticity, and cross-sectional dependency problems in the panel. I know that I can deal with heteroscedasticity and autocorrelation problems using xtreg, "depvar" "varlist", fe vce(robust).
I've also checked Driscol Kraay. It's known that I have to use FGLS method with xtgls STATA command. I've used the following command:
xtgls "depvar" "varlist", panels(hetero) corr(psar1) force
Is that enough to use the command above to remove autocorrelation, heteroscedasticity, and cross-sectional dependency problems simultaneously? if not how can I solve this problem?
I'll be glad if you can help.
Best,
After Hausman test when I run fixed effects model, I have autocorrelation, heteroscedasticity, and cross-sectional dependency problems in the panel. I know that I can deal with heteroscedasticity and autocorrelation problems using xtreg, "depvar" "varlist", fe vce(robust).
I've also checked Driscol Kraay. It's known that I have to use FGLS method with xtgls STATA command. I've used the following command:
xtgls "depvar" "varlist", panels(hetero) corr(psar1) force
Is that enough to use the command above to remove autocorrelation, heteroscedasticity, and cross-sectional dependency problems simultaneously? if not how can I solve this problem?
I'll be glad if you can help.
Best,
Comment