R: feglm 虚拟编码

Question

I´m trying to run fixed effects regressions in R using the feglm command from the fixest package . The data contains three dummy variables: (1) variable z is 0, (2) variable z is between 1 and 8, and (3) variable z is larger than 8. I understand that it is necessary to drop one dummy to avoid collinearity. However, is this done by feglm or do I have to do it manually?

The question boils down to whether I have to include 2 or 3 dummies in my code.

Answer 1

The manual describes a feglm parameter collin.tol :

collin.tol Numeric scalar, default is 1e-10. Threshold deciding when variables should be considered collinear and subsequently removed from the estimation. Higher values means more variables will be removed (if there is presence of collinearity). One signal of presence of collinearity is t-stats that are extremely low (for instance when t-stats < 1e-3).

... so you will have to increase collin.tol to make feglm drop variables for collinearity "on its own accord" (unless their respective t-statistic were indeed even smaller than the catch-all default 1e-10).