Abstract

In order to compile a price index from scanner data, one needs to specify the "product". The product is the unit for which a price and a quantity are observed and which corresponds to the input of the index compilation. Product specification has been recognized as a critical step that can have a significant impact on the results obtained with multilateral methods. Tightly specified products may cause a bias because there may not be sufficient product matches over time. Broadly specified products may cause a bias because there may be quality differences between the underlying transactions that make up the product. In order to assess the biases that result from using either a tight or a broad product specification, we consider the imputation Fisher index as the target index, and compare this index to matched and hybrid Fisher indices. We also define an alternative target index that we call a quality adjusted hybrid Fisher index. We then extend the analysis defined in the bilateral context to the multilateral context of GEKS-type indices. The approach is illustrated on scanner data sets for clothing and hygienic products. This type of analysis should eventually help the compiler deciding which index formula and which product specification to use in practice.