Hello everyone.

I'm doing my final thesis for university and I have a problem I don't know how to solve.

I'm doing a panel data model. My tutor told me to just use FE/RE method as my thesis is quite simple. He told me to do some econometrics tests so my thesis is a bit more complete.

Doing the Hausman test, I obtain that the FE model is the best one to use. After this, I wanted to check about autocorrelation and heteroskedasticity.

For the autocorrelation test I obtained the following results using the command of xtserial

Wooldridge test for autocorrelation in panel data

H0: no first-order autocorrelation

F( 1, 16) = 5.864

Prob > F = 0.0277

And for the heteroskedasticity using xttest3


Modified Wald test for groupwise heteroskedasticity

in fixed effect regression model

H0: sigma(i)^2 = sigma^2 for all i

chi2 (17) = 215.83

Prob>chi2 = 0.0000

I believe to just have a problem of heteroskedasticity. I looked for commands to fix this but all I could find was commands to fix both heteroskedasticity and autocorrelation. Does anyone now how to fix this?

Thank you.