Weighted Averages of Individual Measures of Uncertainty from the European Central Bank’s Survey of Professional Forecasters

Abstract

This paper explores to what extent aggregate measures of uncertainty calculated with data from the European Central Bank’s Survey of Professional Forecasters change when higher weights are given to the data submitted by more accurate forecasters. It is found that these weighted measures suggest higher levels of uncertainty than those obtained with unweighted methods. This result comes from the finding that forecasters with better scores are characterised by higher levels of individual uncertainty, a feature that is robust to the use of different scoring rules and different measures of uncertainty. It is thereby suggested that the European Central Bank could use weighted measures of aggregate uncertainty to complement the assessment obtained from standard unweighted methods.