There are (were?) some issues with F-statistics estimation in stata for IV regressions. We simulate data to reproduce the problem and present a fix in the julia FixedEffectModels.jl library.