Small Sample Bias in Geometric Mean Seasoned CPI Component Indexes

Robert McClelland and Marshall B. Reinsdorf

Abstract

In this analysis we compare the finite sample bias (aka small sample bias) of the geometric means index and the 'seasoned modified Laspeyres index ' (hereafter referred to as the seasoned index.) Our tentative conclusion is that the geometric means index has a larger, growing rate of small sample bias while the seasoned index has a smaller, fluctuating bias.