[Dirah Lestari]

Dear Carlo,
Noted.

Please advise on my next question. By using xtreg, fe robust, the model is free from endogeneity problem?

[Carlo Lazzaro]

Dirah:
it is free from endogeneity as far as time invariant predictors are concerned (as they are wiped out ny the -fe- estimator).
This does not necessarily hold for time varying predictors (see Example #1, -xtivreg- entry, Stata .pdf manual).

[Dirah Lestari]

Dear Carlo,

Please advise on my question. What is the difference between vce(robust) and vce(cluster firm). Can i user vce(robust) to overcome the problem of heteroscedasticity and autocorrelation in the model ?

Thank you

Best regard,
Dirah

[Carlo Lazzaro]

Dirah:
unlike -regress-, -xtreg- calls the very same clustered robust standard error via -vce(cluster panelid)-; -vce(robust)- or simply -robust-.
Clustered robust standard error handles autocorrelation and/or heteroskedasticity.