Academic journal article Real Estate Economics

Rational Expectations, Market Fundamentals and Housing Price Volatility

Academic journal article Real Estate Economics

Rational Expectations, Market Fundamentals and Housing Price Volatility

Article excerpt

In recent years many local housing markets in North America have undergone boom and bust cycles. These episodes are characterized by dramatic house price fluctuations over relatively short periods of time. What drives these episodes of house price volatility? "This is one of the central questions of housing economics," (Mankiw and Weil 1992). As Shiller (1989) asks for asset prices in general: "...can we trace the sources of movements back in a logical manner to fundamental shocks affecting the economy...? Or are price movements due to changes in opinion or psychology, that is, changes in confidence, speculative enthusiasm ... that are best thought of coming ultimately from peoples minds?"

This paper addresses the first question Shiller (1989) poses; can market fundamentals alone explain local house price dynamics? It therefore also provides indirect evidence on the role of psychology in house price cycles. To accomplish this, it derives an asset-based housing price model that explicitly incorporates the hypothesis of rational house price expectations and tests its ability to explain observed short-run fluctuations in real house prices. A number of recent empirical studies find that house price expectations may not be rational, in the sense that information available in one period can be used to predict house prices in future periods. Hamilton and Schwab (1985), Case and Shiller (1989, 1990), Mankiw and Weil (1989), Hosios and Pesando (1991) and Meese and Wallace (1994) all report that house price movements are positively correlated over the short-run. More importantly, information on housing market fundamentals and past price increases can be employed to forecast future excess returns. Mankiw and Weil (1989) simulate the effect of demographic factors on real United States house prices and find that a model with myopic expectations does a much better job of matching observed house price dynamics than does an equivalent model with rational or forward-looking expectations. This implies that the price of housing does not anticipate predictable changes in demand. These findings are inconsistent with rational expectations or semi-strong market efficiency. In an efficient asset market all information relevant to predicting future price is capitalized into current price and excess returns are unforecastable.

Previous empirical work concentrates on testing the implications of rational expectations for single-period return measures. Variants on the standard approach to testing for stock market efficiency are employed. The one-period ex post return to housing investment minus the interest rate is regressed on a constant and a group of other variables known at the beginning of the period. Under the joint null hypothesis of rational expectations and the correctly specified model, coefficient estimates on the additional variables should not be statistically significant. The true ex post return to owner-occupied housing is the sum of capital gains and imputed rental income, which is the dividend from owner-occupied housing. But, because imputed rental income is unobservable, previous researchers employ various proxies based on indices derived from rents on rental properties. Statistical problems inherent in such proxies may be partly driving the rejections of the rational expectations hypothesis since tests based on single period returns require an estimate of actual rents. In addition, standard efficiency tests provide little information about the causes of house price fluctuations.

While true rents are not observed it may be possible to derive the stochastic process they follow. To accomplish this, a method of proxying rents is derived in terms of a parsimonious model of housing market fundamentals. The model for imputed rents is combined with a housing market arbitrage condition to derive a present value model for real house prices. Together with time series models of the stochastic processes driving the fundamental variables, the present value relation yields a reduced form model that relates short-term housing price dynamics to market fundamentals. …

Search by... Author
Show... All Results Primary Sources Peer-reviewed

Oops!

An unknown error has occurred. Please click the button below to reload the page. If the problem persists, please try again in a little while.