Exploring small areas techniques to address uncertainty in Spatial Price Indexes

Abstract

The availability of scanner data for the compilation of price statistics has increased over the past twenty years and several European Member States have introduced Scanner Data into Consumer Price Index (CPI) production. Besides reducing administrative burden, Scanner Data have proved to be of benefit to CPIs thanks to the higher granularity, the wide coverage, the opportunity to implement superlative index and greater precision or lower variance. However, in spite of their potential, to the authors’ knowledge, only few National Statistical Institutes have started official research project for computing sub-national spatial price indexes (SPIs) using Scanner Data. Given the crucial role of SPIs for comparing standard of living across a country it is also relevant to be able to assess their accuracy. In this study, we explore the use of small area estimation techniques to reduce the uncertainty associated to point estimates of sub-national SPIs which we have previously computed via Jackknife Repeated Replications. The data that we use is part of the ISTAT 2018 Scanner Data on the ten provinces of Tuscany (Italy) for selected groups of products.