The predictive performance of multi-level models of housing submarkets: a comparative analysis

Abstract

Much of the housing submarket literature has focused on establishing methods that allow the partitioning of data into distinct market segments. This paper seeks to move the focus on to the question of how best to model submarkets once they have been identified. It focuses on evaluating effectiveness of multi-level models as a technique for modelling submarkets. The paper uses data on housing transactions from Perth, Western Australia, to develop and compare three competing submarket modelling strategies. Model one consists of a citywide "benchmark", model two provides a series of submarket-specific hedonic estimates (this is the ‘industry standard’) and models three and four provide two variants on the multi-level model (differentiated by variation in the degrees of spatial granularity embedded in the model structure). The results suggest that greater granularity enhances performance, although improvements in predictive accuracy will not necessarily offer compelling grounds for the adoption of the multi-level approach.