I am trying to add a did structure to the feols() but after implementing:
model_1 <- feols(y~ treat*post+controls-1 | country+years|0| country, data = dataset)
the did estimator spills out NAs. Does anyone know why this could be the case?
Many thanks.