Dear all, greetings from Paris ! I am doing some panel estimation and I have an issue regarding first-differencing estimations and dummies. I use first difference and gmm estimations for evaluating the impact of a reform dummy occurring through time in a panel of countries, and would simply like to know whether I should differentiate those dummies. The same problem arises when considering country fixed effects that disappear when introducing those variables. Is there any common way to deal with such situation? Any answer on what is typically done in the literature would be very helpful. Many thanks in advance for advice ! Sincerely Franck