From the historians' point of view these assumptions must be realistic or they are junk. If we use the assumption of perfect foresight by businessmen to construct data, the question of its empirical validity is crucial. Altering the assumptions, whether about the model or about the data, can make a substantial difference to both the data and the answers.