This special report from J.P. Morgan’s Economic Research team, which can also be found on J.P. Morgan Markets, offers a look at the housing market correction risk across the United States.
It has been unusual for economic expansions to last more than a few years after the economy reaches full employment, which it now has, and the country should remain on the lookout for early signs of imbalances whose reversal eventually could trigger the next recession. With the recent events of 2007-2009 still lingering on investors’ minds and nominal housing prices now exceeding their peaks from 2006, the housing market is an obvious place to look for signs of overheating.
In this research report, we highlight specific pockets of overheating in detailed data on U.S. counties. For example, high prices, despite expanding supply, such as in Denver, Seattle, Washington, D.C., Portland, Oregon, and Boston indicate some correction risk in J.P. Morgan’s models. Cities with staggering high prices, such as New York City and the San Francisco Bay Area, suggest some chance of correction even with constrained supply. But there is currently no place with the same combination of rapid price growth, rapid debt growth, and expanding supply seen in some areas in 2006.
The nationwide nominal house price index is now 40% above its 2012 low-point and 4% above the peak reached in 2006.

Where the prices are higher
After robust gains over the past five years, the nationwide nominal house price index is now 40% above its 2012 low-point and 4% above the peak reached in 2006. If 2006 was a historic bubble, then current price levels should be looked at more closely. In examining a selection of the largest counties in the U.S., today and back in 2006, there is wide variation across counties both in the level of housing prices and in their change since 2006.
The growth in mortgage debt from 2003 to 2006 explains about half of the variation across counties in the decline in real prices from 2006 to 2011.
Median home values per square foot in Manhattan are about 20 times higher than those in places such as Detroit and Cleveland, whereas the prices in the San Francisco Bay Area are around 10 times higher. While prices in most counties are still below 2006 levels, prices in a handful of places, such as New York City, the San Francisco Bay Area, Boston, Seattle, Denver, and Portland, Oregon have substantially increased since then.
Wide variation across counties exists in the level of housing prices and in their change since 2006.

Where are we now?
Today’s housing market looks quite different from that in the mid-2000s. As households spent much of the recovery paying down debt, there are few places where real debt per capita has increased over the past few years. Even in states such as New York and California that have high and increasing house prices, there has been no sign recently of the kind of debt growth seen during the boom.
While households spent much of the recovery paying down debt, real debt growth per capita in select states has remained positive.
Mortgage debt per capita in select states

Today, unlike in the mid-2000s, there are few counties where prices are high, even though the supply of new housing has increased significantly. The vast majority of counties in the nation that are clustered on the left in the “Mortgage debt per capita in select states” chart feature either low prices and sluggish construction – such as many Midwestern and rust belt cities – or constraints on the supply response to high prices, such as much of the East Coast and California. On the lower right side of the chart, an ample increase in supply has kept prices in check, despite growing demand in southern states such as North Carolina, Georgia, Florida, and Texas. The curve shows the combinations of prices and permit issuance that would produce the same forecast probability of a housing correction as Phoenix in 2006, according to this model. Very few counties fall above or even close to the curve: Places such as New York City and the San Francisco Bay area with staggeringly high prices suggest some chance of correction even with constrained supply, as does elevated prices and substantial supply growth in Seattle (WAKI), Denver (CODN), DC, and Arlington.
‘High prices are less correlated with mortgage debt and more concentrated in supply-constrained areas.’
Where is there greater risk of a correction than in Phoenix in 2006?

What are the risk factors that predict the likelihood of a housing price correction?
The map below identifies a few different variables that would have had some success in forecasting the likelihood of a housing price correction across counties after the boom, and lists the counties currently most at risk according to those factors. Areas such as central California and Las Vegas that have experienced rapid prices gains over the last five years should continue to be monitored.

Conclusion
All housing markets are local, and county-level data do not capture all of the nuances that could drive the price of a given home in a given neighborhood. On the whole, the U.S. housing market still looks considerably less risky than it did in the mid-2000s. While there are some pockets of rapid price growth and extremely high price levels, in addition to a few places with fairly high prices despite growing supply, there is nowhere that has combined these price patterns with rapid debt growth, as occurred in some places in the mid-2000s.

>> BACK TO THE NEWSLETTER: Click here to read other articles from this week’s newsletter

Originally published by J.P Morgan