We analyse individual professional forecasters' beliefs concerning the persistence of GDP shocks. Despite substantial apparent heterogeneity in perceptions, with around one half of the sample of professional forecasters believing shocks do not have permanent effects, we show that these apparent differences may be largely due to short samples and survey respondents being active at different times. When we control for these effects, using a bootstrap, we formally do not reject the null that individuals' long-horizon expectations are interchangeable at a given point in time. When we apply the same bootstrap approach to their medium-term expectations, we do reject the null. We explore this difference between long and medium-horizon forecasts by decomposing revisions in forecasts into permanent and transitory components.
