A0924
Title: Inflation forecasting in persistent times: Revisiting the role of aggregation
Authors: Catalina Martinez Hernandez - European Central Bank (Germany) [presenting]
Abstract: The question of disaggregation is revisited when forecasting headline, core, and food HICP inflation in the euro area. Precisely, it is investigated whether it is better to forecast these rates directly or via aggregation of their subcomponent forecasts. In a real-time forecasting analysis using a state-of-the-art Bayesian vector autoregression, it is found that indeed the breakdowns of headline rate based on five and thirteen special aggregates provide more accurate forecasts, especially for core and food inflation. The inclusion of additional variables in the model only provides additional gains for the forecasts of headline inflation at horizons longer than six months.
