![IconGems-Print](images/IconGems-Print.png)
167 – 167 - Data Mining and Econometrics
Improving Weight Representivity of Fixed Quantity Consumer Price Index Products
Joshua Klick
Bureau of Labor Statistics
The U.S. Bureau of Labor Statistics Consumer Price Index (CPI) program revises fixed quantity weights for products such as the CPI-U, CPI-W, and preliminary C-CPI-U on a biennial basis- each January of even years. The current biennial weight reference period lags from 2-3 years until first use for even index years, and from 3-4 years for odd index years. Reducing this lag via annual revisions improves the timeliness of the fixed quantity weights for index estimation due to the representativeness of household expenditures from a more recent reference period of the Consumer Expenditure Survey. Additionally, annual revisions create a consistent lag across index years, in contrast to the current biennial revision process. Historically, biennial weight revisions were justified in terms of a sufficient sample from the biennial period, and effectively mitigating risk of chain drift. The ensuing analysis will demonstrate that the elementary item-area cell coverage of post-processed annual weights is adequate, and that annual weight revisions pose no appreciable risk of chain drift.