Comment: Binary Regressors in Nonlinear Panel-Data Models with Fixed Effects