ABSTRACT
Over the past two decades modelling activities and model use disappeared from universities to become more and more concentrated at policy institutions. Mankiw (1990) argues that this is mainly caused by recent developments in macroeconomic theory, which have not been of the sort that can be easily adopted in macroeconomic policy models. Another explanation is that practitioners and academics each exploit their own comparative advantages (Fase et al., 1992). Moreover, in policy analysis and forecasting it is common practice to incorporate relevant non-model or judgemental information4 Such information is likely to be available sooner in policy institutions than elsewhere. Whatever the cause of the shift is, it is clear that policy institutions still see macroeconomic models as major information sources and appropriate devices for structuring information in a coherent way.
