Dear Statalist users,

I am estimating a Linear Fixed Effect model in a panel context, which I set as:
Code:
xtset pid wave
where pid is the id of the observed individuals over a panel of 18 waves.

I would like to correct for heteroskedasticity among households and this is why I put vce(cluster hid) where hid is the household id.
Since the panel is over pid I don´t want to change it in xtset.

Code:
xtreg JobS_overall Age Age_sq i.Married Log_RealPay i.Q1-Q3, fe vce(cluster hid)
Stata gives me this message: panels are not nested within clusters

I looked for answer to this problem in the forum but I could not apply them to my problem.
Can someone please explain if there is another way or my purpose is conceptually wrong?

Thank you.

Francesco