Revision of the Hedonic Quality Adjustment Method in the Corporate Goods Price Index and the Services Producer Price Index
October 30, 2020
Research and Statistics Department
Bank of Japan
The Research and Statistics Department (RSD) of the Bank of Japan applies the hedonic quality adjustment method in the compilation of the Corporate Goods Price Index (CGPI) and the Services Producer Price Index (SPPI). The hedonic quality adjustment method is not only a highly objective quality adjustment method, but as the relationship between price and characteristic are estimated statistically, is also applicable to a broad range of changes in product characteristics. Consequently, the hedonic quality adjustment method has been applied to nine products including, for example, IT-related products and passenger cars.
The RSD has reviewed the hedonic quality adjustment method, and in order to deal with multicollinearity and the omitted variable problem -- both of which are well known to exist -- has decided to employ a new estimation method using sparse estimation for the hedonic regression models. The new estimation method allows the RSD to quantify a larger number of price and characteristic relationships through the increased number of explanatory variables used in the regression models. It will also improve the fit of regression models to actual prices as a whole.
In response to the revision of the estimation method and recent changes in market structure -- such as a leveling off or a reduction of the market size and an extension in product life cycle -- the RSD has also decided to terminate the application of the hedonic quality adjustment to several products, to integrate part of the hedonic regression models and to decrease the frequency of re-estimation of the models for some other products.
With these revisions, the RSD intends to enhance the price index and improve its accuracy while facilitating more efficient compilation of statistics. The RSD would like to start to apply these revisions to the CGPI and the SPPI from December 2020.
Please contact below in advance to request permission when reproducing or copying the content of this paper for commercial purposes.
Please credit the source when reproducing or copying the content of this paper.
Price Statistics Division, Research and Statistics Department
E-mail : email@example.com