The Margin of Safety and Turning Points in House Prices: Observations from Three Developed Markets

Using quarterly data from 1960 (United Kingdom), 1963 (United States), and 1977 (Japan) through the second quarter (Q2) of 2010 (all three markets), the authors examined long-run mean-reverting relationships between house prices and inflation, disposable income, GDP, and rents. At the end of Q2 2010, U.S. prices were below their mean-reverting levels and at the lower end of their historical range. Equivalent U.K. and Japanese prices were at or slightly above their mean-reverting levels. Using quarterly data from 1960 (for the United Kingdom), 1963 (for the United States), and 1977 (for Japan, the Tokyo area) through Q2 2010 (all three markets), we examined long-run mean-reverting relationships between new house prices and inflation, disposable income, GDP, and rents. After adjusting for estimated drift and trend coefficients, we found that over the period Q4 2006–Q2 2010, on the one hand, U.S. house prices moved from near the top of their historical range toward the bottom relative to inflation, disposable income, and GDP per capita. On the other hand, yields moved from near the bottom to near the top of their historical range. By Q2 2010, the U.K. market was at trend in relation to inflation and disposable income but slightly above trend in relation to GDP per capita and yields. Japanese new house prices were slightly above trend in relation to inflation, disposable income, and GDP per capita but below trend relative to yields. Hence, for investors and households contemplating the purchase of new homes at the end of Q2 2010, the margin of safety was greatest in the U.S. market, close to historical averages in the United Kingdom, and slightly below average in Japan. In terms of rational expectations theory, U.S. prices were factoring in a lot of pessimism regarding future income and per capita GDP growth, U.K. prices were neither particularly pessimistic nor particularly optimistic, and Japanese prices appeared to be factoring in mildly optimistic assumptions concerning income and economic growth.Although prices in the United States and the United Kingdom appeared high by historical standards immediately prior to the 2008 financial crisis, the precrisis expansion was relatively moderate compared with the expansion experienced in Japan between Q4 1990 and Q2 1991. In fact, Japanese prices peaked at 79 percent, 62 percent, 58 percent, and 30 percent away from the whole-sample-period trend values for, respectively, inflation, disposable income, GDP, and yields. These percentages can be compared with equivalent whole-sample deviations of 17 percent, 15 percent, 13 percent, and 21 percent for the U.S. market and 33 percent, 36 percent, 34 percent, and 45 percent for the U.K. market.A comparison of the three markets for new houses shows that the U.S. market has the strongest mean-reverting tendencies relative to trends and also the most statistically significant positive and negative trend coefficients. Furthermore, deviations from trend appear to have a smaller range and standard deviation in the U.S. market than in the other two markets for all the deflated house price series and yields. The reason may be that the U.S. market faces fewer supply constraints than the other two markets, where population density is much greater; hence, the relatively greater price elasticity of supply in the U.S. market has a more moderating effect on price expansions relative to the underlying macroeconomy than it does in Japan and the United Kingdom.To speculate that house prices are to rise or fall indefinitely is to speculate that per capita wealth will also rise or fall indefinitely. Given technological advances, this assumption may not be unrealistic, but history has shown that long-term trends in economic growth are punctuated by cyclical expansions and contractions. For most people, housing investment provides a geared (levered) but, in the medium term, imperfectly correlated exposure to economic growth. Neglect of this final observation is liable to result in a failure to provide an adequate margin of safety when making decisions on housing choice, investment, and housing policy. Despite apparent extremes of relative over- and undervaluation, no instances occurred in the past sample periods when prices failed to revert to either the rolling or the whole-sample trend in any of the three markets studied.