An economic indicator faces two requirements. It should be timely reported and should not significantly be altered afterward to avoid erroneous messages. At the same time they should reflect changing market conditions constantly and appropriately. These requirements are particularly challenging for housing price indices, since housing markets are subject to large temporal/seasonal changes and occasional structural changes. In this study we estimate a hedonic price index of previously-owned condominiums of Tokyo 23 Wards from 1986 through 2006, taking account of seasonal sample selection biases and structural changes in a way it enables us to report the indexes timely which are not subject to change after reporting. Specifically, we propose an overlapping-period hedonic model (OPHM), in which a hedonic price index is calculated every month based on data in the “window” of a year ending this month (this month and previous eleven months). We also estimate hedonic housing price indexes under alternative assumptions: (i) no structural change (“structurally restricted”) and (ii) different structure for every moth (“structurally unrestricted”). Results suggest that the structure of the housing market, including seasonality, changes over time, and these changes occur continuously over time. It is also demonstrated that structurally restricted indices that do not account for structural changes involve a large time lag compared with indices that do account for structural changes during periods with significant price fluctuations.
清水 et al. (Mon,) studied this question.