Special Studies, April 29, 2008

Paul Emrath, Ph.D.

Report available to the public as a courtesy of HousingEconomics.com

Table of Contents
Print the Full Article
Table 1
Table 2
Table 3
Foreclosures, Production, Prices, and Homeownership Rates by state (Excel tables below) are available for Subscribers only
Northeast
Midwest
South
West
See previous Special Studies

Over the past year or two many local housing markets have experienced trends with which stakeholders have had little prior experience. The trends vary from state to state, but include rising foreclosure rates, declining production and prices of owner-occupied housing, and a falling rate of homeownership.

In California, for example, the foreclosure start rate was 1.04 percent for all loans and 4.62 percent for subprime loans in the fourth quarter of 2007. These numbers were up about 158 percent and 146 percent, respectively, from the fourth quarter of 2006. Single family permits, meanwhile, were issued at a (seasonally adjusted annual) rate of 47,912—a historically low number for California and down 43 percent from the 84,493 posted in the fourth quarter of 2006 (which itself was down more than 40 percent from the fourth quarter of 2005). The OFHEO purchase only index indicates that house prices in California declined 11.5 percent between the fourth quarter of 2006 and the fourth quarter of 2007, and the homeownership rate in the state dropped 2.3 percentage points, from an even 60.0 percent down to 57.7, over the same period.

Attached tables show this information for all 50 states and the District of Columbia from the first quarter of 2000 through the fourth quarter of 2007. There are four Excel files, one for each of the four principle Census regions: Northeast, Midwest, South, and West. If you are interested in foreclosure start rates, single family permits, house prices, or homeownership for a particular state over this time span, open the file for the appropriate region and move to the worksheet for a particular state by navigating across the tabs that appear at the bottom of the screen. Note these tables are for subscribers only .

Foreclosures and Production

A natural question to ask, based on these data, is whether or not there is any particular relationship among the variables. Do any of them move together in a predictable pattern?  One way to get a rough initial picture of this is by calculating simple “correlation coefficients” that lie between +1 and -1 for the variables two at a time. [1] Table 1 shows the correlations between housing and foreclosure variables that tended to be strongest across most of the states. The most powerful relationship found is the one between the foreclosure start rate (the percent of loans entering the foreclosure process during a given period) and single family permits. [2] The relationship is negative: when foreclosures rise, single family permits tend to decline.

This is generally what we would generally expect. The most immediate impact from a rise in foreclosures is an increase in existing homes on the market and competing with new construction for customers. This is one factor that will tend to drive new production down, and the correlation is one way to measure the strength of this effect. In addition, the cause for an increase in foreclosures is an indication of distress in the market beyond the added vacant homes. The most frequently cited causes for increased foreclosures in this time period have been falling house prices, unaffordable mortgage loans and rising unemployment. Falling house prices will also chill new production as builders attempt to reprice existing inventory and find ways to reduce costs to compete at lower price points. Falling house prices can also cause mortgage rates and availability to tighten, which eliminates some buyers and reduces demand. Rising unemployment also reduces demand as households see declines in their income. Hence, the rise in foreclosures has both a supply and a demand impact on the housing market and one reaction is lower production.

Although total permits (including multifamily) and single family permits both tend to go down when foreclosures rise, the effect is somewhat stronger on single family permits. Again, this is not surprising. Increased foreclosures also tend to depress homeownership, leading to reduced single family production in particular, while the market for multifamily rental properties can actually benefit from this to a certain extent. (A substantial part of the market for new multifamily properties consists of condominiums, however, and this segment of the market tends suffer from any effect that reduces single family production.)

The negative relationship between subprime foreclosures and single family permits is generally strong and persistent. Nevertheless, there is a lot variation among the states. The subprime foreclosure/permit correlation is actually positive in three states (Alaska, Maryland, and Iowa), although never higher than .11. The negative relationship between total foreclosures and permits is not quite as persistent (the correlation is positive for 13 states). However, the list of states with the strongest total foreclosure/permit correlations (between -.81 and -.90) include states where the anecdotal evidence of problems in housing and mortgage markets has been quite strong: Michigan, California, Massachusetts, and Nevada. Tables 2 and 3 contain similar information to Table 1. Table 1 lists the states in alphabetic order, Table 2 sorts them by the total foreclosure/permit correlations, and Table 3 sorts them in descending order of the subprime foreclosure/permit correlations.

House Prices

Correlations of various factors with house prices tend to be so variable that it’s better to look at the experiences of states one at a time than to attempt to draw general conclusions across states. [3] For example, there are a number of states (including large states like New York, Florida, and Texas) where house prices and the homeownership rate show a strong positive correlation. But there are also several states (such as Virginia and North Dakota) where homeownership and house prices have tended to move in opposite directions.

When studying the relationship between house prices and foreclosures, stronger  correlations were usually found by using house prices lagged by four quarters. In other words, a change in house prices tended to be followed by a change in foreclosures a year later. Caution is advisable when interpreting this relationship, however, because a correlation—even with the lag—doesn’t necessarily prove that rising house prices were causing foreclosures. Both could ultimately result from underlying factors (relaxed underwriting standards or a change in the mix of mortgage products) operating in particular housing markets.

Moreover, the correlation between lagged house prices and the total foreclosure rate is actually negative in thirteen states plus the District of Columbia, and as low as -.82 in Utah. The correlations are generally less strong between lagged house prices and subprime foreclosures, with similar numbers of negative and positive correlations, as well as correlations that are less extreme in both directions. Many states show relatively similar rankings by both measures. Colorado, for instance, ranks first with the highest correlations of house prices with both total and subprime foreclosures (+.97 and +.72, respectively); while Utah ranks at the bottom by both measures (-.82 and -.63). But there are counterexamples like Indiana, a state that combines a lagged house price/total foreclosure correlation of +.88 with a lagged house price/subprime foreclosure correlation of -.17.

There also is a relationship between house prices and single family permits four quarters later. In about two-thirds of the cases this relationship is negative—i.e., a change in house prices tends to be followed by a change in permits in the opposite direction a year later.

Foreclosures and Homeownership

In about two-thirds of the states, there is a negative relationship between the total foreclosure rate and the home ownership rate four quarters later. In other words, in these states, when foreclosures rise, homeownership declines about a year later. This agrees with intuition, as foreclosures imply that some households are moving out of owner-occupied homes.  Moreover, they are usually moving out under financially distressing circumstances that makes another home purchase in the immediately future unlikely. Hence, some owners become renters and the homeownership rate declines.

It is somewhat difficult to explain counter examples like Texas and Virginia, where increases in foreclosures tend to be followed by increases in homeownership (and the correlations are as high as .74 and .68, respectively). It may be that there are other factors at work in these states that the simple correlation fails to capture, or that the timing is of the effects is somewhat different in these states. In Virginia, the total foreclosure rate surged the last two quarters of 2007 while the homeownership rate plummeted. In Texas, the movements of both the homeownership and total foreclosure rates have been comparatively mild.

Compared to the correlations between total foreclosures and homeownership, the correlations between subprime foreclosures and homeownership are similar, but generally  weaker. Again, this agrees with intuition. Because total foreclosures impact a larger fraction of home owners than subprime foreclosures, it seems reasonable that the effect of total foreclosures on homeownership would be stronger.

Conclusion

Across the 50 states, with very few exceptions, the number of single family permits issued has declined substantially since 2005. In many (but not all) states, this reduction has been accompanied by other signs of distress such as declining homeownership, depreciating house prices, and loans entering the foreclosure process at an increased rate.

Although simple pairwise correlations between these variables don’t explain state housing markets as fully as more elaborate statistical models, they may be useful as a first attempt to isolate some of the relationships. The strongest relationship found persistently across  states is the tendency for the rate of single family permits issuance to decline as foreclosure rates rise. Other strong correlations were often found, but the state-to-state variation proved to be extreme. For this reason, it would be wise to avoid general conclusions for the country as a whole and focus instead on the pairwise correlations among foreclosures, production, house prices, and homeownership one state at a time.

Regions:

Northeast: Connecticut, Maine, Massachusetts, New Hampshire, New Jersey, New York,
Pennsylvania, Rhode Island, Vermont.

Midwest: Illinois, Indiana, Iowa, Kansas, Michigan, Minnesota, Missouri,  Nebraska, Nor th Dakota, Ohio, South Dakota, Wisconsin.

South: Alabama, Arkansas,  Delaware, District of Columbia, Florida, Georgia, Kentucky,  Louisiana, Maryland, Mississippi, North Carolina, Oklahoma, South Carolina, Tennessee, Texas, Virginia, West Virginia.

West: Alaska, Arizona, California, Colorado, Hawaii, Idaho, Montana, Nevada, New Mexico, Oregon, Utah, Washington, Wyoming.

Return to the Top 

____________________________

Footnotes:

[1] When the correlation between two variables is positive, it means that they tend to move up or down together. When negative, it means when one goes up, the other tends to go down. A coefficient of 1 means that two variables are “perfectly” correlated.

[2] The study began by looking at total permits as well as single family permits. The correlations tended to be similar in either case, but generally stronger when the single family measure was used (the exception being states like New York, where an unusually large share of residential production tends to be multifamily).

[3] The measure of house prices used for this purpose comes from the U.S. Office of Federal Housing Enterprise Oversight (OFHEO). It is a repeat-transactions index based on mortgages purchased or securitized by Fannie Mae and Freddie Mac. As such, it includes only mortgages that conform to Fannie/Freddie underwriting standards, but has the advantage of uniform coverage across all 50 states. The “purchase only” OFHEO index is used, which excludes mortgages used to refinance properties.