search

sergiocorreia / reghdfe

Linear, IV and GMM Regressions With Any Number of Fixed Effects
http://scorreia.com/software/reghdfe/
MIT License
212 stars 56 forks source link

empirical Bayes shrinkage after reghdfe #199

Open laszlosandor opened 4 years ago

laszlosandor commented 4 years ago

What would it take to do shrinkage on fixed effects after reghdfe, absorb( , savefe), as briefly discussed here (with further links therein)?

sergiocorreia commented 4 years ago

It would not be completely trivial.

Reading the online appendix, they make some assumptions that seem extremely strong ("the standard error is our estimate of the standard deviation of the asymptotic distribution") and at a quick glance don't make much sense (if many groups only have a few obs. then the estimated SE will be quite biased away from zero.

After that, I guess it should be pretty independent of reghdfe, so you could code it as a separate ado.