This paper proposes a simple method to compute prediction intervals for expert-adjusted forecasts in case the analyst does not have the underlying model forecasts and thus has to create own approximate model forecasts, based on data available to the analyst. An illustration to airline revenues data shows that experts can substantially reduce forecast uncertainty.

Airline revenues, Approximate model forecasts, Expert-adjusted forecasts, Forecast uncertainty, Prediction intervals
Time-Series Models; Dynamic Quantile Regressions (jel C22), Model Evaluation and Testing (jel C52), Forecasting and Other Model Applications (jel C53)
hdl.handle.net/1765/130345
Advances in Decision Sciences
Department of Econometrics

Franses, Ph.H.B.F. (2018). Prediction intervals for expert-adjusted forecasts. Advances in Decision Sciences (Vol. 22). Retrieved from http://hdl.handle.net/1765/130345