Berkeley Program on Housing and Urban Policy

A random walk in time and independence in space are maintained hypotheses in traditional empirical models of housing prices. However, there is increasing evidence in the context of hedonic models that housing prices are predictable over time and space. This paper examines the price discovery process in individual dwellings by relaxing both assumptions, using a unique body of data from the Singapore private condominium market in a repeat sales framework. We develop a formal model that tests directly the hypotheses that the prices of individual dwellings follow a random walk over time and that the price of an individual dwelling is independent of the price of a neighboring dwelling. The empirical results clearly support mean reversion in housing prices and also diffusion of innovations over space. This predictability may suggest that excess returns are possible. When aggregate returns are computed from models that assume a random walk and spatial independence, we find that they are strongly autocorrelated. However, when they are calculated from models permitting mean reversion and spatial autocorrelation, predictability in investment returns is completely absent. Despite this, an extensive simulation of investor performance, over different time horizons and with different investment rules, indicates quite clearly that recognition of the spatial and autocorrelated nature of prices substantially improves investor returns. The magnitude of deviations from standard models of price dynamics are small, but their economic implications are quite large in the housing market.