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

Abstract

Item-level transactions data yield cost-of-living indices that can account for quality change and consumer substitution. Transactions data require confronting the rapid turnover of items because prices of new and existing products are interrelated in equilibrium. This paper evaluates multiple approaches to measuring quality change at scale. It shows that a hedonic superlative approach—using econometrics or machine learning for hedonic estimation combined with index formulas that require simultaneous observation of item-level price and expenditure—yields improved measures of the cost of living. Accounting for ubiquitous quality change and for consumer substitution yields lower measures of inflation than traditional, official methods.