The construction of House Price Indexes (HPI) for Australian capital cities poses particular difficulties. Because the same houses are rarely sold in successive quarters, conventional price methods based on matched samples cannot be applied. Furthermore houses are heterogeneous goods in terms of location and other characteristics. Quarterly changes in price indexes could therefore be driven by compositional change rather than underlying price movements.
Hedonic methods involve application of regression techniques to express the price of a home in terms of its characteristics (such as age, size, construction material, location, neighbourhood etc.) and then constructing an index from the hedonic function. This research indicated that plausible hedonic models could be fitted for two cities.
Readers should note that since the completion of this research, ABS has revised the methods for compiling the house price index as described in ABS cat. no. 6417.0. While ABS has not adopted hedonic methods due to the lack of data available across the capital cities, they nevertheless remain of interest in ongoing research and have been used indirectly to inform the development of the new methods.