I am asked from time to time to contribute to a group of people who forecast national house prices. A lot of me would like to help (after all, who doesn't like being asked for his opinion), but I am never sure how to answer, because to me, there is no "national" house price.
To see why, take a look at the Case-Shiller Index for 20 United States metropolitan areas from 2000 until now. Take percentage changes in the index, and then find the correlations across metropolitan areas. If metropolitan areas moved in lock-step with each other, they would have correlations of one. If they have nothing do to with each other, they will have correlations of zero.
Some metropolitan areas do have strong correlations with each other--Seattle and Portland, for example, have a correlation of .84. But others, such as Dallas and Las Vegas (at .30), or Charlotte and Cleveland (at .23), have very low correlations with each other, meaning that knowing what is happening in one market tells us little about what is happening in another (anyone who wants the correlation matrix is welcome to email me at richarkg@usc.edu).
This matters, because the impact of house prices on the national economy is asymmetric. Suppose if house prices all moved together. Then if national prices were to rise by 4 percent, house prices everywhere would move by 4 percent, and we could almost dismiss default as a possibility. But house prices rising 13 percent in, say Boston, might happen at the same time that house prices are falling by 5 percent in, say, Atlanta (that is not a forecast, by-the-way). The average of the two cities is plus four percent, but while we would expect to see few if any defaults in Boston, we would likely see some in Atlanta, as borrowers with small down-payments would go underwater. Falling house prices, whether regionally or (in the recent but unusual case) nationally tend to push up defaults, which in turn puts pressure on financial institutions, which in turn affects the entire economy. When the rust belt economy suffered in the late 1970s, house prices there fell, and we saw lots of default. The same was true of Texas in the middle 1980s, New England in the later 1980s and the early 1990s, and California for a good part of the 1990s. These were periods of pretty heavy defaults, and yet national house prices rose every year.
