Working Paper: CEPR ID: DP2499
Authors: Paolo Giordani; Paul Sderlind
Abstract: We study the inflation uncertainty reported by individual forecasters in the Survey of Professional Forecasters 1969-1999. Three popular methods of estimating uncertainty from survey data are analysed in the context of models for forecasting and asset pricing. We find that inflation uncertainty fluctuates over time in a way that traditional time series models fail to capture. Instead, uncertainty is highly correlated with the level of inflation, in particular with recent positive inflation surprises. We also find that disagreement among forecasters increases with the inflation rate and causes above-average fluctuations in individual uncertainty.
Keywords: survey data; survey of professional forecasters; TGARCH; VAR
JEL Codes: C53; E31; E37
Edges that are evidenced by causal inference methods are in orange, and the rest are in light blue.
| Cause | Effect |
|---|---|
| inflation uncertainty fluctuates over time (E31) | inflation uncertainty (E31) |
| inflation uncertainty is significantly correlated with the level of inflation (E31) | inflation uncertainty (E31) |
| positive inflation surprises (E31) | inflation uncertainty (E31) |
| inflation rate (E31) | disagreement among forecasters (E17) |
| disagreement among forecasters (E17) | above-average fluctuations in individual uncertainty (D89) |
| rising inflation (E31) | increased uncertainty (D89) |
| negative inflation surprises (E31) | inflation uncertainty (E31) |
| new method of decomposing uncertainty (D80) | average individual uncertainty and disagreement (D80) |
| average individual uncertainty and disagreement (D80) | behavior of financial markets (G41) |