Lags for Newey West Standard Errors in fmb

[Prateek Bedi]

Hi,

I am using the command fmb in order to estimate Fama-Macbeth regressions for my unbalanced panel dataset. I would like to use the Newey-West standard errors for my model which the fmb command supports. However, I am not sure as to how should I calculate the appropriate lag length for Newey-West standard errors. Any guidance is highly appreciated.

Thanks!

[Prateek Bedi]

Hi,

I am facing this problem again. I need to specify a lag length for calculating Newey-West standard errors as part of the fmb option in the command called asreg. Also, I would like to know when should we use 'Newet-West' standard errors instead of the conventional OLS standard errors. I am working on an unbalanced panel dataset comprising 1696 firms and 16 years of data. The residuals exhibit serial correlation and heteroscedasticity.

Any help is highly appreciated!

[Prateek Bedi]

Hi,

After searching some older posts on Statalist, I found two rules of thumb for deciding lags for Newey-West standard errors. These rules of thumb are mentioned in post #4 (thanks to Andrew Musau) of the following thread.

https://www.statalist.org/forums/for...lation-command

In addition, the following links may also be useful:

https://www.bauer.uh.edu/rsusmel/phd/ec1-13.pdf

http://econweb.umd.edu/~chao/Teachin...Estimation.pdf

Thanks!