Clustered Standard Errors

[Lucy Smith]

Hi,

I am using Clustered Standard Errors for a fixed effects regression.
I know that they control for serial correlation within clusters, and the same for heteroskedasticty. Is this the same for normality as well?

Thank you.

[Joao Santos Silva]

Lucy,

If you have enough clusters, you do not need normality to perform valid inference.

Joao