I have panel data. I employ fe model, but there are heteroskedasticity and serial correlation problems in model. How i can solve this problem? using robust standard errors is apcetable? which model is in such case optimal?
Yes you can follow the robust estimation method. For more details please read page 38 of below attachment.
https://www.princeton.edu/~otorres/Panel101.pdf.
In case you use Stata (but not restricted to it), have a look at https://www.statalist.org/forums/forum/general-stata-discussion/general/1475615-newey-regression-for-panel-data
Hello Levani Gafrindashvili. You can adopt Driscoll and Kraay’s method. It addresses both problems. Find attached their paper. The method is applicable in Stata.
Regards.
in case u use STATA u need to use "cluster" option to fix such issues of heteroskedasticity and serial correlation
See this reference -> Article Robust Standard Errors for Panel Regressions With Cross-Sect...
- Badges
- Science topic
- Similar topics
- Social Psychology
- Attitude