Bias Correction in a Stable AD(1,1) Model