correcting for heteroscedasticity

[Tina Samuel]

Hi all,
I am running a panel regression with exporter importer and country pair fixed effects using the reghdfe command. I tested for heteroscedasticity using white test and Breusch-Pagan test and both indicated that there is heteroscedasticity. I want to correct for that but I am not sure how, my main dependent variable is the log of exports, is there's another way I can correct for heteroscedasticity other than log transformation? I can run a PPML regression instead but i still need to include the exporter, importer and country pair fixed effects.
would greatly appreciate your input on this.

[Joao Santos Silva]

Dear Tina Samuel,

Just estimate the model by PPML using the user-contributed ppmlhdfe command.

Best wishes,

Joao

[Tina Samuel]

Yes, I used the PPML and it worked really well. Thank you for the suggestion. Also, I tested for autocorrelation using the Wooldridge test (using PPML model) and it indicated that there's a serial correlation. I used the cluster option to fix it; however, I am wondering if the cluster option is enough or do I need to test-again after using the cluster option?
Also if there's a better option to use than cluster that would be great.

Thanks.

[Tina Samuel]

Yes, I used the PPML and it worked really well. Thank you for the suggestion.

Thanks.

[Joao Santos Silva]

Clustering should be fine.