A Granular Approach to Density Forecasting Euro Area Inflation

Inflation forecasts play a key role in guiding decisions of households, governments, firms, and central banks. The sharp rise in European inflation after the pandemic and the Russian invasion of Ukraine highlighted the need for rethinking the conventional linear inflation forecasting models that cannot capture non-linear dynamics. While most forecasters rely on aggregate measures of inflation and other macroeconomic variables, recent research shows evidence that using disaggregated inflation series (which includes components from gasoline to cinema tickets) improves the point forecast of aggregate inflation. Our research aims to contribute to this strand of literature by focusing on density forecasting of the euro area aggregate inflation. We combine the predictive densities of each component, accounting for the uncertainty around their forecasts, using the Bayesian Predictive Synthesis (BPS) technique, proposed by McAlinn and West (2019).