Stark New Warning on Home Prices

Two very different recent reports, one factual and one theoretical, recently address the same issue; are home prices too damn high?  And if so, how will we know.

The theoretical discussion, part of Freddie Mac's most recent Insights and Outlook report looks at different ways of assessing affordability and overvaluation.  The actual may be a little more alarming.

In an article on Bloomberg.com, Prashant Gopal says that buyers' tolerance for high prices may be reaching a limit in some big U.S. cities. Using tools developed by Allan Weiss, principal of Weiss Residential Research LLC and a co-creator of the Case-Shiller home prices indices, Gopal reports that almost half of the single family homes in the Washington, DC and New York metropolitan areas appear to be declining in value.

Weiss has developed a repeat sales index of homes including their sales and appraisal history, location, and size along with the valuation and sales histories of a large set of similar houses that have sold at least twice over a period of time.

According to the Bloomberg report, the values of 45 percent of houses in the two cities dropped by at least 2 percent in June 2015 compared to a year earlier.  In June 2014 only 15 percent of homes in Washington and 20 percent in New York lost value.  The index also noted declines in larger numbers of properties in other cities; Los Angeles, Chicago Phoenix, and Miami.

Gopal says rising home prices coupled with little growth in incomes and tight credit are drying up the pool of buyers, especially as cash buyers - investors and foreign buyers - are no longer as numerous. 

Weiss told Bloomberg that, while an average home in a city may be rising, others could be losing value.  A growing share of such homes could be an early warning that the market may be in danger.  "If you have a market where every house is rising and you hear the news that the housing market is up, you're correct in applying it to your own home," Weiss said. "However, if you are in a market where 60 percent of houses are rising, you have a 40 percent chance of misunderstanding what's going on with your house."

Freddie Mac's economists saying a house is overvalued can indicate a lack of affordability, a credit risk that will ultimately lead to delinquencies and defaults or it might be in anticipation that prices are likely to fall at some future point.  As each of these symptoms of overvaluation have implications for public policy and the economy so the housing industry has devised a number of affordability statistics that measure the level and trend and the reasonableness and sustainability of house prices. 

Common sense would indicate that house prices should bear some relationship to area incomes and the simplest measure is the ratio of the median house price in an area to the median household income such as these measures of Houston and San Francisco.

Median income in San Francisco has been relatively high over this 22 year period, about 1.4 times the median income in Houston, still house prices in San Francisco have averaged 4.8 times the median compared to only 2.6 times the median in Houston.  But Freddie Mac says this comparison can't really tell us whether homes are overvalued in the California city.  For one thing, it tells us nothing about the amenities and services residents enjoy and that contribute to home values.

One way to control for the value of location is to compare the price/income ratio to its average value over time.  When these normalized ratios are plotted for the two cities, the long-term median price-to-income ratio in San Francisco is 3.8 but in the first quarter of 2015 it was 6.0, 58 percent higher suggesting those houses are 58 percent overvalued. The same exercise for Houston reveals the possibility prices are 15 percent too high.

This approach, however, is highly sensitive to small changes in the affordability statistics and if a long term average is used rather than a long-term median San Francisco appears to be only 18 percent overvalued and Houston is actually undervalued by 16 percent.

Another method is to compare the monthly mortgage payment on the median priced house to the median monthly income, an approach that adds the impact of higher or lower interest rates.  Other analysts compare monthly mortgage payments to median rents in the area. Both are good measures of affordability but require additional normalization to address overvaluation.

Exhibit 3 gives visual insight to the confusion that can result from these all reasonable measures.  It displays March 2015 estimates of value for ten metro areas.  Two models are represented along with the simple price-to-income ratio normalized by both the long-run median and the long-run average ratio. In half of the areas the various methods come to quite opposite conclusions.

Exhibit 4 provides another illustration of the range of answers provided by these metrics. The vertical axis displays the estimates of overvaluation provided by comparing the current-price-to-income ratio to the long-term median ratio.  The horizontal axis displays a statistical model estimate.  Again the difference between the approaches makes it difficult to form definite conclusions about many of the metro areas.

And what of national prices?  While we have recent experience with high prices; from 2003 to 2007 national home prices rose 8.6 percent annually and 39 percent over the whole period.  In retrospect we can see those increases were unsustainable and the correction inevitable but at the time analysts didn't think so.  By 2006 most people agreed that the rate of increase couldn't be sustained but could not agree on the future price path.

There are many examples of local house price collapses due to unique situations within an area.  But overall U.S. house prices had not declined since the Great Depression, and many observers were skeptical that this event could ever be repeated. Only a handful of investors were sure enough of their analyses to bet on a nationwide housing collapse.

Freddie Mac says there is regrettably no single statistic that identifies when a housing market is overvalued or when prices are likely to fall.  Substantial human judgment is also required and at best these affordability statistics can only wake us up to potential danger.

Or they could ask Allan Weiss