William Miles
Final draft: December 2017
International Journal of Housing Markets and Analysis
Home prices and their fundamental determinants, such as income, should, according to theory and intuition have a stable long run relationship. While bubbles exist, these are deviations from this long run relationship. However, numerous empirical studies have failed to find such long run stationarity between home values and income. This presents a puzzle-could house prices really drift away from income indefinitely, with no tendency to return to some equilibrium? Previous studies have imposed linear adjustment to deviations from equilibrium in their tests. However, it has been clearly established that both home prices and income exhibit nonlinear dynamics. Moreover, tests for stationarity have low power in the presence of nonlinearity. We accordingly address this issue for the G-7 countries. First, we have a longer span of data than previous studies. In addition we employ tests which detect and allow for nonlinear adjustment to shocks. For the US, we are able to reject the null hypothesis of nonstationarity in the price/income ratio with this longer span of data and the use of a linear, but powerful testing procedure heretofore not utilized in this literature. However, we do not find a stable relationship using this procedure for the remaining six G-7 nations. We do find evidence, for these six countries, of nonlinearity, and upon applying a test which posits asymmetric adjustment we obtain results which indicate stationary, long run relationships between values and income in five of the remaining six G-7 countries.