Proposition of the Quality Adjusted Geks-Type Price Index

Jacek Białek, Natalia Pawelec
Statistika, 105(3): 291–305
https://doi.org/10.54694/stat.2024.71

Abstract
Scanner data from retailers like supermarkets, electronics stores, and online shops provides detailed transaction information at the barcode level (e.g., GTIN, EAN), allowing for the use of various price index formulas, including weighted ones. Due to high product turnover and seasonality, multilateral index methods are ideal, as they use a whole-time window and are transitive, avoiding chain drift. However, most commonly used multilateral indices (e.g., GEKS, CCDI, GK, TPD) fail the identity test, which requires the index to return to one when prices revert to their original levels. This paper proposes a new multilateral index inspired by GEKS but incorporating quality adjustments like the Geary-Khamis method. The index satisfies the identity test and other key axioms, demonstrating its robustness. Comparisons with the SPQ index and quality-adjusted indices (GEKS-AQU, GEKS-AQI) confirm its effectiveness, making it a highly useful tool for scanner data analysis in both theory and practice.

Keywords
Scanner data, multilateral indices, the GEKS-index