A review of bias in the CPI

Abstract

This paper discusses one of the high profile issues in economic statistics in recent years. The possible existence of bias in consumer price indices (CPI) became an international measurement issue following the release of the Boskin Report in the United States in 1996 although many of the issues had been under investigation for some time. The research before and after the Boskin Report has been widespread and undertaken by many different analysts. This paper attempts to draw together the discussion surrounding bias and to determine appropriate definitions and interpretations of the issues involved. The main conclusion is that it is essential that CPI compilers and users continually assess the index being compiled to ensure its relevance to the economic environment it is attempting to explain.