Interpreting and evaluating CESIfo's World Economic Survey directional forecasts

Using the Carlson and Parkin (1975) framework and employing the Pesaran–Timmermann (1992) Predictive Failure statistic, we evaluate several consensus forecast series from CESIfo's World Economic Survey. Several issues are examined related to interpreting qualitative survey responses. We define what an “about the same” response implies across different economic variables, the value of agreement across the forecast panel, and how to maximize the signal value provided by the survey. We find that survey respondents provide statistically significant directional forecasts or signals.