February 17, 2005


U.S. Home Prices: Does Bust Always Follow Boom? (FDIC FYI, February 10, 2005)

U.S. home prices have boomed in recent years. Average U.S. home prices rose 13 percent in the year ending September 2004, and are up almost 50 percent over five years. In December 2004, the Office of Federal Housing Enterprise Oversight (OFHEO) noted, "The growth in home prices over the past year surpasses any increase in 25 years."1 Because of this rapid growth, some have become concerned about the possibility of a home price collapse, either nationwide or in a number of major cities.

But before we evaluate the implications of the recent housing boom, it is useful to put it in a historical context. How extensive has the surge in home prices been in recent years? What can history tell us about the likelihood of "booms" to go "bust"? This issue of FYI examines the historical movement of home prices at the metro level to gain insight into the outlook for U.S. home prices. There is some evidence that home price booms can be followed by busts—although we find, at least by our criteria, that this pattern may be more the exception than the rule.

In order to examine the historical evidence of home price booms and busts, we first need to arrive at some definition of a "boom." Although there are many possible ways to approach this issue, we chose a fairly simple definition based on a cursory examination of cities that have exhibited some of the strongest home price cycles in recent decades. We define a "boom" simply as a 30 percent or greater increase in inflation-adjusted (or real) home prices during any three-year period. For our "1/3 in 3" rule, we adjust the nominal home price series that is published by OFHEO using the Bureau of Labor Statistics consumer price index (CPI) less the price of shelter, which is used by OFHEO to adjust home price changes for inflation.

Table 1 summarizes our findings. It shows that applying our "1/3 in 3" rule results in the identification of a number of individual metro-area price booms since 1978. In fact, 63 different U.S. metropolitan areas have experienced at least one boom during that period, and 24 cities experienced more than one boom. Geographically, home price booms have been concentrated in cities in California and the Northeast, which account for almost 70 percent of our 63 boom markets. This share may be overstated, however, due to the limited availability of data for many cities outside these areas prior to 1990.

Defining a Housing "Bust"
One way to measure home price busts in our historical sample would be to start with our definition of a boom (real price increase greater than 30 percent in 3 years) and simply reverse the sign and look for price declines. However, applying this approach proves to be too stringent a definition, resulting in the identification of only five metro-area price busts since 1978. The reason this measure proves to be too stringent is that home prices tend to adjust slowly (or be "sticky downward," in economists' terms) during a downturn. Unless homeowners have lost the means to maintain their mortgage payments, say through mass layoffs, or are forced to move due to some other circumstance, they typically have the option to withdraw their homes from the market—especially if they feel the price being offered by potential buyers is too low. Because prices are sticky downward, it will be necessary to define a price bust using a lower threshold and a longer time period, such as a real price decline of 15 percent or more in five years.

Applying a "15 in 5" definition of declining metro-area home prices, we find 142 metro areas where the average home price, adjusted for inflation, has declined by at least 15 percent over a five-year period. But is this definition now too lenient, resulting in the identification of too many cities? After all, what we are really saying is that the value of the average owner's home in these 142 metro areas simply failed to keep up with inflation during the five-year period and fell behind inflation by at least 15 percent. The price of their home in nominal terms may never have fallen at all. For example, the five-year change in the CPI less shelter index between 1978 and 1982 was 43 percent. A city such as Akron, Ohio, where homeowners saw the value of their homes rise by 12 percent during this period, would nonetheless be placed in the "bust" column under a "15 in 5" definition based on changes in real home prices.

A period of true distress for homeowners and lenders might be better defined in terms of a large decline in nominal prices. Since mortgage debt is taken on and paid off in nominal dollars, a decline of more than 15 percent in nominal home values could push the value of many properties below what homeowners owe on their mortgages. If homeowners had no choice but to sell in this type of situation, they could be forced to bring a sizeable personal check to the closing. Such a large decline in nominal home values would reduce the incentive of homeowners to repay their mortgages to protect their equity stake, since equity tends to evaporate with a decline in prices. This is why we feel that a better measure of distress in metro-area housing markets would be to define a bust as an average decline in nominal home prices of at least 15 percent over five years, or a nominal "15 in 5" rule.

Using our criteria, some 21 cities can be defined as having experienced a housing bust at some point during the past 25 years. Table 1 highlights these cities in red and shows two major episodes of home price busts.3 The first began in the mid-1980s in the "oil patch" cities of Texas, Oklahoma, Louisiana, and some of the western states. This episode includes the most severe price declines of our entire sample, with nominal prices in one city falling by as much as 40 percent over a five-year period. Another episode of large nominal price declines occurred in selected metro areas of the Northeast and California beginning in the early 1990s.4 Other cities that met our criterion but were not associated with these two major episodes included Peoria during the mid-1980s and Honolulu, where nominal prices declined for six straight years through 2001.

Before going further and analyzing the historical evidence for booms gone bust, it should be noted that most U.S. cities have demonstrated fairly stable home price trends over time. [...]

Although this paper demonstrates that relatively few metro area housing booms have ended in busts, there are reasons to think that history might be an imperfect guide to the present situation.

Boy, you have to go through an awful lot of contortions to make bewlieve there's even a potential problem.

Posted by Orrin Judd at February 17, 2005 7:35 AM

The widely shared growth of wealth poses a real problem for the old Left. The common man was not supposed to be able to acquire any, and policies aimed at confiscating the wealth of those few who did manage provided swag for programs and ideological moral preening.Now, everybody wants to get in on the (wealth) act and the Left is still standing there ready to swipe the boon. Who but a lunatic would adhere to a "property is theft" jag?

Posted by: Luciferous at February 17, 2005 1:06 PM

All property markets are local. Out here where there is a plentiful supply of developable land, House prices rise very gently. Costal area, especially in Blue states where nimbyism runs rampant and developable land is hard to find, run the risk of bubbles.

Posted by: Robert Schwartz at February 17, 2005 10:18 PM

If interest rates go up, the price of houses will go down, or flatten for a very long time. People buy houses based on how large a payment they can carry, not on the price of the house. It's just that simple.

A return to Carter era interest rates will crash the housing market and probably the banking system along with it.

Posted by: Bart at February 18, 2005 11:00 AM