U.S. flag

An official website of the United States government

Department of Labor Logo United States Department of Labor
Dot gov

The .gov means it's official.
Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you're on a federal government site.

Https

The site is secure.
The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

Bureau of Labor Statistics > Office of Survey Methods Research > Publications > Browse Research Papers

Combining Hedonic with Multilateral Indexes for Turnover and Chain Drift in Transactions Data Consumer Price Indexes

Gregory Kurtzon

Abstract

This paper combines hedonics and rolling window multilateral indexes in transactions data to fix the problems of product turnover and chain drift. Transactions data hold the potential to eliminate substitution bias in elementary price indexes with quantities concurrent with prices, but using these quantities can cause chain drift (non-circularity). The typical solution in national consumer price indexes (CPIs) is to use multilateral indexes, but these don’t fully eliminate chain drift. The large amount of product entry and exit is usually delt with by making unit value prices, but this causes biases from quality changes and Jensen’s inequality. Previous research has studied different multilateral formulas, window lengths, and rolling window splicing methods, and noted that sales and product turnover are highly connected to chain drift. Therefore, hedonic imputation should help all these issues at the same time. I create Pakes (2003) style hedonic indexes using IRI grocery and drug store data from 2001-2011 with dummy variables for all variable values plus interactions and bootstrap the indexes to check for overfitting. I make Caves, Christensen, and Diewert (1982) (CCDI) indexes over rolling window lengths that span the entire range of data and make Multiperiod Identity Tests and Laspeyres-Paasche spreads to evaluate them. Chain drift is typically improved for short rolling windows, but doesn’t disappear, and can actually get better or worse for longer windows until the window length converges to all twelve years. The hedonic indexes have an acceptable level of variance and their use to some extent has the potential to reduce chain drift themselves by smoothing the effects of sales.