Hedonic Price Indexes and the Matched Models Approach

We consider three approaches to estimating quality‐adjusted price changes: (i) the dummy variable approach from a hedonic regression, (ii) a superlative or exact hedonic index and (iii) a matching technique—a technique akin to that used by statistical offices. The dummy variable approach is prevalent in the literature and has been used for independent estimates of quality changes when commenting on sources of error in consumer price indexes. However, the availability of scanner data provides an opportunity to utilize data on the prices (unit values), volumes and quality characteristics of a much wider range of transactions and to consider methods less restrictive than the dummy variable approach. The practical use of superlative or exact hedonic index and matching techniques using scanner data is explored, and the results from all three methods are compared. A feature of the paper is the breadth of the empirical work. It not only encompasses three different approaches, but extends across four different types of consumer durables. The manner in which the three approaches relate to each other is explored and the implications for quality‐adjusted price changes is discussed.