Quantile regression with fixed effects and clustered standard errors

[Mathilde Lesueur]

Dear all,

Here is the context: I observe individuals within cells and at two time periods. I am running a regression with cells and time fixed effects, giving me coefficients denoting within-cell variation of individual characteristics across time. Standard errors are clustered at the cell level.

Now, I would like to run a quantile regression with cell and year fixed effects and cluster standard errors. However, from what I read:

• qreg and qreg2 do not allow fixed effects but allow to cluster se
• xtqreg and qregpd allow fixed effects but not to cluster se.

Is there a way, using one of these commands, to run the quantile regression I intend to?

Any help is greatly appreciated. Thank you very much in advance!

Mathilde

[Joao Santos Silva]

Dear Mathilde Lesueur,

If I understand correctly your situation, there is no QR estimator with FE that would work. I suggest you use correlated random effects.

Best wishes,

Joao