This paper presents an AR(1) model in the spirit of Hausman and Taylor zero differences between actual and in-sample predicted trade flows. Large systematic differences between observed and in-sample predicted trade flows only indicate model misspecification and econometric problems....