Do house prices experience periodic bull and bear markets like the stock market? Or are they stable in real (inflation-adjusted) terms most of the time, with big disruptions once or twice in a century? Two popular house price series tell these very different stories. Knowing which is better will lead to superior investment outcomes and improved policy decisions.
Karl (Chip) Case, of Wellesley College, and the Nobel Prize-winning Yale professor Robert Shiller, have constructed the most widely-known suite of indices, which are now part of the S&P index family. Here is the Case-Shiller national house price index in real terms from 1890 through December 2013:
Case-Shiller National House Price Index in Constant Dollars, 1890-2013
And here is a house price series distributed by the data firm of Crandall, Pierce & Co., consisting of the median new home sales price in constant dollars collected by the U.S. Department of Housing and Urban Development. (For brevity, we call this the “Crandall” series.)
Crandall, Pierce Median New Home Sale Prices in Constant Dollars, March 1963-March 2014.03
Source: Crandall, Pierce & Co., Libertyville, IL. Reprinted with permission.
Could any two charts describing the same underlying phenomenon look more different? In the Case-Shiller chart, there was one great bear market in the last 50 years, from late 2006 to early 2012, following a massive price expansion or bubble.1
In the Crandall chart, however, bull and bear markets have alternated in a remarkably regular pattern. All of the bear markets represent losses of roughly 20%, with the crash of 2008-2011 only a little worse than the three other housing bear markets that occurred in 1968-1970, 1979-1982, and 1988-1992. The Crandall chart also shows real prices rising pretty smartly – 1.35% per year – while the Case-Shiller chart shows a much slower rise.
Note that the two price series do not purport to measure the same thing. The Crandall data are for new houses only; the Case-Shiller data are intended to reflect the entire stock of housing capital.2 The Crandall data are for a median house, the size and quality of which are constantly changing; Case and Shiller explicitly adjust for changes in the size and quality of a house. There are many other differences, so it’s understandable that the two series disagree somewhat – but they’re both intended to track house prices, so the contrast between them is striking and troubling.