Quality Adjustment at Scale: Hedonic vs. Exact Demand-Based Price Indices

Abstract

This paper explores methods for constructing price indices from item-level transactions data on prices, quantities, and product attributes. The paper evaluates approaches that are feasible at scale, i.e., across the wide range of products, disparate encoding of attributes, and rapid product turnover inherent in “big data” on economic transactions, while producing improved cost-of-living indices that reflect both substitution effects and quality change. The paper presents hedonic methods that estimate changing valuations of both observable and unobservable characteristics in the presence of product turnover. It also considers demand-based methods that account for product turnover and changing appeal of continuing products. The paper provides evidence of substantial quality-adjustment in prices for a wide range of goods, including food and high-tech consumer products. The paper also shows that hedonics can be implemented with well-encoded attributes using standard econometrics and with unstructured attribute data using machine learning.