ABSTRACT
Based on the UK case, we can start with the distinction made between what were apparently the two tasks of the UK Panel of Independent Forecasters (in spite of its name), namely to provide forecasts of economic development in the UK and to formulate policy prescriptions (see Evans, 1997, Chapter 13 in this volume). With respect to the use of models in the forecasting process the discussions of the Panel appear to focus on the role of judgement in the forecasts. On the issue of modelbased policy analysis, their discussions involved the differences in opinion on the working of economic mechanisms, or the type of theories incorporated into the models. In the case of policy prescriptions, the statement of Patrick Minford in the discussions in the Panel (see Evans, 1997, Chapter 13 in this volume) makes it clear that here there is a third distinction, namely between policy analysis, based on mechanisms at work in the economy, and recommendations based on policy preferences (e.g. about the allowable size of the sacrifice ratio). An even clearer example of this distinction is the use of the Dutch CPB model in calculating the economic programmes of the political parties at election time. Here the same model is used so that the assumed mechanisms are constant, but the policy preferences are those expressed in party programmes.
