Empirical Bayes Shrinkage Estimator
This Stata .ado program implements the Empirical Bayes estimator of Morris (1983) and is described in Appendix C of Chandra, Finkelstein, Sacarny, and Syverson (2013). The program’s usage is described in the file itself. If you are performing shrinkage on fixed effects you will also need estimates of their standard errors; see the program below to generate those estimates.
The Empirical Bayes method is useful whenever you want to use a noisily estimated variable as a regressor. Generally in the presence of noise (i.e. measurement error) regression estimates are biased. However, when the variance of that noise is known, this method can adjust (shrink) the variable to restore unbiasedness to the regression.
Fixed Effect Standard Error Estimator
This Stata .ado program, based on fese by Austin Nichols, estimates standard errors for the fixed effects in a linear regression. The standard errors will match what Stata would report if you included the fixed effects as dummies without an omitted category and no regression constant. The program is optimized for homoscedastic standard errors and will return them very quickly. Adding heteroscedasticity-robust standard errors makes it quite slow, and this option is not as well tested as the one for homoscedastic standard errors, either. See the file itself for usage. Note that Austin Nichols asks that you cite him if you use his program in your papers and this program is a derivative of his.