This paper examines Lionel Robbins' critical attitude towards formal empirical work from the standpoint of modern econometrics. It argues that his attitude towards empirical work rested on indefensible assumptions and that he failed to realise that the role he saw for empirical work undermined his belief in the primacy of economic theory. This matters because Robbins' attitudes are echoed in modern economics, best exemplified by the calibration methodology of Kydland and Prescott, which is vulnerable to similar criticisms.