Do real housing prices revert to some trend? If so, where do they stand now with respect to trend? In their February 2010 paper entitled “The Margin of Safety and House Price Turning Points: Observations from the US, the UK and Japan”, Mitsuru Mizuno and Isaac Tabner investigate real housing price deviation from and reversion to trend in three developed markets. Using quarterly measures of housing price, inflation, disposable income, GDP and rent from 1960 (UK), 1963 (U.S.) and 1977 (Japan) through 2009, they conclude that:
- Housing prices adjusted for inflation, disposable income and GDP appear to revert to trend (but not at a predictable pace) in all three countries. Reversion to trend in the U.S. housing market appears to be the strongest.
- Real housing prices in the U.S. / UK / Japan (deflated by respective consumer price indexes) appreciate at annualized rates of 1.9% / 5.5% /1.9% over the sample periods.
- Maximum upside (downside) deviations of real housing prices from these trends in the U.S. / UK / Japan are 17% / 33% / 78% (-16% / -30% / -27%).
respectively. The maximum upside (downside) deviation from trend in the U.S. occurs in the first quarter of 2007 (last quarter of 1970), with the 2007 peak not particularly extreme. The upside extreme in Japan during the late 1980s and early 1990s far surpasses any observed in the U.S. or the UK over the sample periods. - As of the the third quarter of 2009, the U.S. housing price is at the low end of the observed range, while prices in the UK and Japan are at, or slightly above, their trends.
- Sample periods are very short in terms of number of housing boom-bust cycles.
In the following chart, taken from the paper, the light blue line plots real (CPI-deflated) U.S. housing price starting from a base value of one, and the red line depicts the trend (1.9% annualized growth). The black line is a detrended series computed by dividing the actual series by the trend value. As of the the third quarter of 2009, the U.S. housing price appears to reflect a high level of pessimism.
The paper also presents charts for housing prices similarly deflated by GDP and disposable income, for which differences between trending and detrended results are less pronounced.
In the next chart, also from the paper, the light blue line plots the yield from rent on housing (rent/price), and the red line indicates the trend (-0.1% annualized growth). The black line is the detrended series. The rental yield is an indication of return on investment in housing and of the opportunity cost of owning versus renting.
The paper offers comparable charts for the UK and Japan.
Note that an investor in housing could not have known in real time how far the adjusted housing price for a quarter deviated from the total-sample trend. The investor could have used cumulative to-date trend or rolling historical trend, but such real-time (knowable) trends would often be very different from the total-sample trend.
In summary, evidence from small samples indicates that housing price adjusted for inflation, disposable income and GDP reverts to trend (and that adjusted U.S. housing price is currently low).
