Title: Forecasting US inflation in real-time
Authors: Chad Fulton - Federal Reserve Board of Governors (United States) [presenting]
Kirstin Hubrich - Federal Reserve Board (United States)
Abstract: A real-time forecasting exercise for US inflation is performed, investigating whether and how additional information - additional macroeconomic variables, expert judgement, or forecast combination - can improve forecast accuracy and robustness. This is particularly relevant now, as the length of the current expansion nears that of the longest on record and unemployment has fallen to a rate not seen for nearly sixty years. Distinguishing features of our study include a focus on forecasting performance before, during, and after the global financial crisis and the use of (published) Federal Reserve Board staff forecasts contained in Tealbooks. We find that (1) while simple models remain hard to beat, model-based forecasts of the types we consider can improve on them, especially in the post-crisis period; (2) judgmental forecasts are competitive and can outperform even simple benchmark models; (3) aggregating forecasts of inflation components can improve performance compared to forecasting the aggregate directly; (4) forecast combination approaches provide competitive forecasts and robustify against bad forecasts.