Predicting Inflation: Professional Experts Versus No-Change Forecasts

2010/10/12 by Tilmann Gneiting, Thordis L. Thorarinsdottir, Gneiting, Tilmann +1
Earth and Planetary Sciences · Economics, Econometrics and Finance · Environmental Science · #Applications (stat.AP) #FOS: Computer and information sciences #Hydrology and Drought Analysis #Meteorological Phenomena and Simulations #Monetary Policy and Economic Impact

paper · pdf · doi:10.48550/arxiv.1010.2318

openalex publication_date 2010/10/12 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/28

Abstract

We compare forecasts of United States inflation from the Survey of Professional Forecasters (SPF) to predictions made by simple statistical techniques. In nowcasting, economic expertise is persuasive. When projecting beyond the current quarter, novel yet simplistic probabilistic no-change forecasts are equally competitive. We further interpret surveys as ensembles of forecasts, and show that they can be used similarly to the ways in which ensemble prediction systems have transformed weather forecasting. Then we borrow another idea from weather forecasting, in that we apply statistical techniques to postprocess the SPF forecast, based on experience from the recent past. The foregoing conclusions remain unchanged after survey postprocessing.

Citations

• A real-time data set for macroeconomists

Related

• Practice Makes Perfect: Learning Effects with Household Point and Density Forecasts of Inflation
• Why Has U.S. Inflation Become Harder to Forecast?
• Forecasting US Inflation Using Bayesian Nonparametric Models
• From Fixed‐Event to Fixed‐Horizon Density Forecasts: Obtaining Measures of Multihorizon Uncertainty from Survey Density Forecasts
• Forecasting and Explaining the Phillips Curve: A SHAP-Based Comparison of Machine Learning and Traditional Time-Series Models for Canadian Unemployment and Inflation