“…Recently, this line of research has been taken up again by Joseph, Kalamara, Kapetanios, Potjagailo, and Chakraborty (2022) who use disaggregate inflation data combined with machine learning methods to forecast headline inflation in the UK. Related to this academic debate, central banks have always forecasted different components of the inflation rate, both for statistical reasons and for understanding the underlying price dynamics (Benalal, del Hoyo, Landau, Roma, and Skudelny, 2004;Capistrán, Constandse, and Ramos-Francia, 2010;Huwiler and Kaufmann, 2013;Giannone, Lenza, Momferatou, and Onorante, 2014;Ulgazi and Vertier, 2022). By using the German inflation rate which is the one with the most detailed breakdown worldwide, we show that combining disaggregate inflation nowcasts into an aggregate nowcast for headline inflation is a highly competitive approach.…”