2005
Abstract:
This research unites the two major strands of work that exist to date in the literature on
Housing Markets. The first is the notion of spatial equilibrium wherein consumers
inhabiting different units are thought to be at a constant utility level. As a consequence
prices "compensate" for differential "hedonic" housing attributes. The second is the
application of life-cycle analysis to the determination of the "full" cost of owning
housing as a financial asset. Linking the two we hypothesize that it is the "risk-adjusted
annual cost of ownership" which should compensate owners for the differential
consumption flows that come from various houses. To test whether this is the case we
develop a unique data set for 4 US metropolitan areas that ascertains the appreciation and
risk from owning housing at the ZIP code level. We then combine this with transaction
based data on price levels - at the same level of geographic detail. We find that in ZIP
codes with higher historic appreciation, price levels are indeed higher, but we suspect that
this may represent misspecification through an identity. When we test over a shorter
period for whether prices anticipate future appreciation - we get very mixed results. In
nearly half of the specifications ex ante appreciation are insignificant or have the wrong
sign. The results for risk are similarly disappointing. These results reinforce the doubts
raised by others over whether the housing market is "efficiently" priced.
