Hi everyone! May I ask if it is technically correct to include both current and lagged values of an independent variable in xtabond2? I tried to do it but it gives me coefficients with different signs. The predicted signs should be the same.