Methodological choices when using multilateral methods: window length, seasonality and aggregation structure, Eurostat

Multilateral methods are increasingly being used to derive consumer price indices from transactions data. Eurostat released two guidance publications in 2022 on which the present paper is based. It discusses three key choices to be made when using multilateral methods: 1) the window length over which the multilateral method is compiled and the splicing method used; 2) the compilation methodology for categories that include seasonal products; 3) the aggregation structure and the methodology for integrating diverse data sources in the compilation. With splicing, one cannot avoid chain drift entirely; hence the drift must be reduced as much as possible. Individual seasonal products do not need to be explicitly identified in the data but in order to have a good coverage of two successive in-season periods, it is important that the time window is wide enough. The advantage of applying a multilateral method up to a more aggregated level is the ability to cope with a dynamic universe of individual products. Keeping the variable weights as high as possible in the index hierarchy allows changes in the importance of products and/or outlets over time to be captured.