Retrospective Computations of Price Index Numbers: Theory and Application

Abstract

Due to outdated weighting information, a Laspeyres-based Consumer Price Index (CPI) is prone to accumulating upward bias. Therefore, the present study introduces and examines simple and transparent revision approaches that retrospectively address the source of the bias. They provide a consistent long-run time series of the CPI and they require no additional information. Furthermore, a coherent decomposition of the bias into the contributions of individual product groups is developed. In a case study, the approaches are applied to a Laspeyresbased CPI. The empirical results confirm the theoretical predictions. The proposed revision approaches are not only adoptable to most national CPIs, but also to other price level measures such as the producer price index or the import and export price