Do sales prices overstate underlying house prices in market downturns? Evidence from the Canadian house price crash of 1991

Canadian Journal of Administrative Sciences, Dec 2002 by Marion Steele, Richard Goy

These studies indicate, however, that the upward bias is quite stable. Thus, the change in homeowners' estimates is a good indicator of the true change. This gives some confidence that indexes using homeowners' estimates will yield unbiased estimates of the change in the price of houses in the stock.

Theoretical Issues

What is the True Price?

This paper is concerned with investigating whether the change in the quality-adjusted price of houses in the stock is different from the change in the MLS average price. How do these two prices relate to the true price? Indeed, what is the true price? It could be taken as the steady-state equilibrium price, the one for which there are no mismatches between homeowners' preferences and their houses' characteristics (Wheaton, 1990). Alternatively, the true price could be taken as a price on the equilibrium price path between steady states. Fu (1996) characterizes such a path as depending on the relative arrival rates of buyers and sellers in a world in which consumption and investment demands for housing are not separable. A price on the equilibrium price path is clearly closely related to the price of houses that transact,s while the steady state equilibrium price is closely related to the price of houses in the stock. Thus, equilibrium analysis gives no single answer to the question of what is the true price.

Neither Wheaton (1990) nor Fu (1996) deals directly with the issue of the noisiness in house prices, although it lurks beneath the surface of Fu's model. This noisiness-high variance in price for a house of given characteristics at a given time-results from heterogeneity, search costs, and thin markets. Noisiness means that an observed transaction price may be quite different from the one that a homeowner reasonably expects to realize, because, for example, houses that actually transact may be ones at the top of the distribution of purchasers' bid prices.

Early in a market downturn it is plausible that observed price is higher than expected price, a phenomenon associated with the stylized fact of a large drop in sales volume.9 This drop occurs because homeowners are unwilling to sell at a price close to the mean of the buyers' bid price distribution. Stein (1995) suggests that a reason for the drop in volume in downturns is the reluctance of some mismatched homeowners to sell because of their low or negative equity. Chan (2001) suggests an explanation is vendors' loss aversion, which indeed is a motivation underlying some of the reserve price rules examined by Goetzmann (1996). Fu (1996) suggests that potential buyers with an accumulated mismatch between house characteristics and preferences early in the downturn do not buy because they expect further declines, and the loss in utility resulting from the mismatch is insufficient to offset the net investment benefits of delayed purchase. In our view, the drop in sales volume is exacerbated by transactions risk, the coordination and price risk of a move in an illiquid and falling market. This risk arises fundamentally because of the inseparability of consumption and investment. A homeowner who plans to sell before buying faces the possibility that there will be no house which matches his preferences and is available for purchase at the expected price once he has sold his house. He might face the Hobson's choice of renting a suitable house (and not satisfying his investment demand) or satisfying his investment demand (but not his consumption preferences) by purchasing an unsuitable house. At the same time, a homeowner who plans to buy before selling faces the possibility that, once he has bought, he will not be able to sell his own house except at a price far below the middle of his price distribution.