COICOP2018 preparation and HICP-CT calculation (COPAC)

ended

This project focuses on two important topics for the HICP in the coming years: the switch from the Ecoicop classification to the new Coicop2018 classification and a new method for calculating the HICP at constant tax (HICP-CT). The introduction of the COICOP2018 classification is planned simultaneously with the base year shift of 2025. In order to have a smooth introduction we need to solve classification issues, determine new weights and make modifications to the existing sample in order to properly compile the HICP according to COICOP2018 classification. In addition there is a need for compiling back series with the COICOP2018 classification to accommodate the needs of end users. The main results of this work package are preliminary COICOP2018 back series from 2010 till 2022, a method to convert National Accounts product group expenditures to COICOP2018 sub classes and new classification strategies to properly classify price collection data to COICOP2018. With the introduction of multilateral methods in the HICP calculation, also a discussion about the calculation of the HICP-CT has emerged. Adjusting observed prices (“price-method”) so that tax rates are kept constant is the usual method. Two considerations have to be made when following this approach: detailed information about tax rates at product level is required, which is not always available, and the HICP compilation process is repeated for the HICP-CT, which leads to a duplication of calculation time each month. The second objective of this grant proposal is to investigate the feasibility of calculating the HICP-CT directly on indices at elementary aggregate level (“index-method”). This will be done for the multilateral method that is used at Statistics Netherlands for transaction data, that is, the Geary-Khamis or “QU method”, and the method HASP-25 for extending index series from month to month. If successful, this will simplify and speed up the HICP compilation process, also for traditional methods.