Housing Bubble Concerns and the Outlook for Mortgage Credit Quality
U.S. home prices have risen briskly over the past several years, outpacing growth in disposable income. In terms of sales volumes, the housing sector had another banner year in 2003.1 This housing boom has raised concerns among some analysts about the possibility of a home price bubble and the specter of home prices suddenly collapsing. Rising levels of consumer and mortgage debt, and a decline in the quality of that debt, also merit attention. What connects these areas of concern is the possibility that household credit quality could decline further if home prices were to fall precipitously. This article reviews current evidence and expert opinion on the possibility of a national home price bubble and considers the overall outlook for mortgage credit quality for the remainder of 2004.
Some Market Watchers Perceive a U.S. Housing Bubble...
Karl Case and Robert Shiller define a bubble as "a situation in which excessive public expectations for future price increases cause prices to be temporarily elevated."2 Under this definition, an asset bubble can be said to exist when prices have risen faster than the underlying supply and demand fundamentals would suggest. The speculative element is a key feature of any asset bubble, as expectations of further price gains, rather than fundamental factors, begin to drive appreciation. Typically, these episodes are identified only in hindsight, when panic selling bursts the bubble. This sequence of events has been observed in many historical episodes with assets such as equities, land, and even tulip bulbs. There is an important distinction, therefore, between a "boom," when prices increase at a historically rapid pace, and a "bubble," when unsustainable factors lead to a boom/bust price path. A boom is required for a bubble to form, but a boom does not necessarily mean that a collapse in prices is imminent.
Concerns about a nationwide housing bubble arise from the very robust activity in the housing sector and mortgage markets in recent years. Mortgage borrowing has boomed as low interest rates, together with strong demographics, have spurred homeownership and refinancing by existing homeowners looking to liquidate home equity gains. In fact, the homeownership rate reached an all-time high of 68.6 percent by the fourth quarter of 2003. The mortgage market has grown at double-digit rates since early 2002, bringing total U.S. home mortgage debt to $6.6 trillion by third quarter 2003. In just the two years ending September 2003, total mortgage debt outstanding rose by $1.4 trillion, or 26 percent.3 Strong demand for housing, facilitated by low interest rates, has pushed home prices to their highest rates of appreciation in more than a decade. But this sturdy price appreciation has not been accompanied by equally strong personal income growth. Since 2000, annual home price appreciation has averaged roughly 7 percent, while disposable per capita personal income gained 4 percent per year, on average (see Chart 1). As a result, many observers, still smarting from the high-tech stock bubble of 2000, are uneasy over the longevity of this housing upturn and its seeming disconnect from fundamentals such as income. They fear an abrupt end to what they perceive as a bubble in the value of U.S. homes.
...But Housing May Be More Resistant to Bubbles than Other Assets
On the surface, the performance of the housing market in recent years appears to be consistent with a burgeoning asset bubble. And because housing typically is a leveraged asset, purchasing, or "investing," in housing can be viewed as comparable to buying stocks on margin, with similar risk, in that a homeowner can potentially go "underwater," owing more on the asset than it might fetch if resold at the prevailing market price.
However, the analogy to stocks ends there. There are several reasons why housing, particularly owner-occupied housing, is less prone to price bubbles than stocks. For example, homeowners cannot short their housing asset readily, nor is there a margin call demanding additional funds when prices drop below the outstanding mortgage balance. Also, the trading volume in the housing market is much less than in the financial asset markets, since housing turns over far less frequently.
Further underscoring differences in these assets are unique characteristics and structural attributes that make housing less vulnerable than any other asset to severe boom and bust price swings (see Table 1). Housing's unique characteristics of utility, transaction mechanisms, and financial treatment are not comparable to those of assets such as equities. The confluence of these factors mitigates speculative tendencies and allays the potential for a price collapse.
Attributes of the Housing "Asset" That Mitigate Against Price Collapse following a Boom, in Contrast to Stocks
Housing provides shelter as well as privacy, choice, and comfort. Homeowners live in their "asset." Stockowners do not.
High Transaction Costs
The costs required to secure and vacate a house are huge with respect to time, fees, effort, and household disruption. The high transaction costs for housing generally discourage rapid, repeat buying and selling, unlike equity markets, where discount brokerages have pushed trading fees to very low levels in recent years.
Homeowners enjoy the tax benefit of mortgage interest deductibility, and buyers can usually deduct loan points and origination fees. Alternative living arrangements, such as renting or living with one’s parents, do not convey similar tax relief. Also, unlike stock sales, gains from home sales are tax advantaged in some instances.
Breadth of Ownership
There is no "nationwide" market for housing as there is for other assets, and a large majority of single-family homes are owner-occupied.a Stock ownership is more likely than housing to reflect trader/investor holdings, which are maintained only as long as they are profitable or provide some benefit, such as diversification. Stock prices reflect millions of decisions brought together in a single collective marketplace. Stock investors’ buy-sell decisions in response to reported earnings can immediately and significantly influence stock price movements. In contrast, housing prices are driven by local factors, as most homeowners do not live in homes that are geographically far removed from their jobs or family.
Homeownership is thought to convey other important benefits aside from investment returns and shelter, contributing to neighborhood stability and social involvement. In addition, there is some evidence that homeownership contributes to more positive outcomes for children.b
a Of total occupied single-family units, 84.6 percent are owner-occupied according to the U.S. Bureau of the Census, American Housing Survey for the United States, 2001. b Richard K. Green and Michelle J. White, "Measuring the Benefits of Homeowning: Effects on Children," Journal of Urban Economics, 41, 1996, pp. 441–461.
There Is No U.S. Housing Bubble, but Local Volatility Has Occurred
Although reliable price histories do not exist, it is recognized that home prices fell precipitously in many areas of the nation during the Great Depression. This drop could be likened to a national boom/bust or bubble in housing. However, the structure of mortgage finance, which played an important role in that episode, has changed profoundly since that time. Before the 1930s, mortgage credit typically took the form of short-term, callable, nonamortizing loans. As home prices fell, homeowners often could not refinance when their loans came due or were called by the bank. The result was a wave of real estate liquidation that drove prices downward. In response, federal mortgage programs established in the aftermath of the Depression took the form of long-term, noncallable, amortizing notes that created the standard of modern U.S. mortgage finance. These institutional changes eliminated a major avenue for any systemic price collapse in U.S. housing markets resulting from a severe nationwide economic shock.
While some systemic factors, such as mortgage interest rates, influence home prices across the country, most influences on home prices are local. They include supply-side factors, such as the availability of developable land and local construction costs, and demand-side factors, including changes in employment, real incomes, population, and taxes.4 Differences in these local factors tend to drive differences in home price performance across the country (Chart 2). Markets where home price boom/busts have occurred in the past few decades have tended to exhibit two common features: (1) significant episodes of economic boom and bust, affecting demand through large employment and population changes, and (2) limited space for new development, constraining adjustment in the supply of new housing.
Many areas of the "oil patch" of Texas, Louisiana, and Oklahoma experienced large percentage declines in home prices during the last half of the 1980s. Major markets in New England and California experienced smaller price declines between 1989 and 1995. These regions shared some of the problems associated with the "rolling regional recession" of the 1980s, when boom times were followed by job losses, bank failures, and, often, significant outmigration. On a national scale, however, the U.S. repeat-sales home price index, published by the Office of Federal Housing Enterprise Oversight (OFHEO), has never shown an annual decline in its 27-year history.
Price declines such as those associated with the rolling regional recession
of the 1980s typically can be attributed to demand-side factors. However, supply-side
issues also come into play. While markets typically clear by adjustments in
price and quantity, markets where supply is constrained must clear more through
price change. Coastal markets or those with strict zoning, such as in California
and the Northeast, historically have shown wider swings in home price appreciation,
attributable in part to the difficulty of expanding supply during a boom (see Map
Thus, owing to the combination of demand- and supply-side factors, markets with both volatile demand and constrained supply are likely to see more pronounced price swings than markets experiencing just one or the other. While a nationwide housing bubble appears unlikely at this time, some markets are experiencing significant price swings, reflecting changing local economic, demographic, or affordability conditions.
The most recent OFHEO data (see see Table 2) show that markets registering the weakest home price growth are, for the most part, cities that have seen significant recent economic deterioration as a result of the loss of dot-com or telecom jobs or, in the case of Salt Lake City and Provo, Utah, a post-Olympics slump. The markets with the strongest home price growth are mostly cities in California and the Northeast that typically have shown a tendency toward wide price swings because of supply constraints. These markets generally did not experience a disproportionate level of economic distress during and after the 2001 recession.
Variations in Home Price Appreciation (annual percentage change in home prices, third quarter 2003)
10 Fastest Markets
10 Slowest Markets
Fort Pierce, FL
San Jose, CA
San Diego, CA
Salt Lake City, UT
Fort Collins, CO
Santa Barbara, CA
The history of U.S. home prices suggests a clear potential for home prices to decline in individual markets, particularly in cities that have shown wide price swings in the past and where prices recently have risen dramatically. However, this same history also strongly suggests that it is highly unlikely that home prices will fall precipitously across the entire countryeven if rising interest rates raise the cost of mortgage borrowing and reduce housing affordability. Further, a significant price decline does not inevitably follow a sharp rise in local home prices. In many cases, the aftermath of a housing boom has been characterized by slower sales and price stabilization until the underlying fundamentals have a chance to catch up with market prices.
Understanding the behavior of both buyers and sellers is key to understanding home price dynamics. Were prices to fall in certain markets, history and academic research suggest that potential sellers would tend to withdraw from the marketplace rather than proceed with panic sales. In fact, studies show that "household mobility [selling] is significantly influenced by nominal loss aversion," or a willingness to continue to hold the asset and take further losses in the hope that the price will go up one day.5 Because of homeowners' loss aversion, homes tend to stay on the market longer with asking prices set well above selling prices, and many sellers withdraw their homes without sale.6 This behavior is typical in all but the most economically distressed markets. Unless the number of homeowners who must sell because of job or income loss is a significant portion of sellers in a market, weakness in a local real estate market is more likely to result in a slowdown in transactions than a plunge in home prices. Although owners may be more inclined to sell in a down market if the property is not their primary residence, high transaction costs weigh against quick "trades" by investors.
Credit Quality Concerns May Pose Greater Risks
The previous discussion indicates that concerns over an imminent, widespread collapse in home prices are somewhat misplaced. Of greater and broader concern are rising levels of consumer and mortgage debt and the decline in the quality of that debt. The recent period of historically low mortgage rates has left the economy in uncharted territory with respect to household indebtedness. Among the new homeowners who have pushed the nation's ownership rate to record levels are those with high leverage, volatile incomes, or limited wealth. Many of these buyers realized the American dream of home ownership only through low-down-payment, variable-rate mortgages obtained during a time of historically low interest rates. As interest rates rise, these homeowners may see their incomes strained by rising debt service costs, while slower home price appreciation limits their ability to build equity and lower their leverage. The remainder of this article explores these household credit quality concerns.
Mortgage credit quality indicators exhibited some weakness during the 2001
recession and, in the ensuing period, have shown slight improvement. Since
peaking in the third quarter of 2001 at 4.83 percent, delinquencies on all
residential mortgage loans have declined to 4.28 percent.7 Similarly,
credit quality problems in construction and development (C&D) loans have
moderated somewhat. The ratio of delinquent C&D loans to total C&D
loans has fallen to 1.8 percent as of September 2003 from its recent peak of
2.6 percent two years ago.8 However, volatility
in home prices could contribute to consumer credit quality concerns. According
to data from LoanPerformance Corporation, more than three-quarters
of currently outstanding mortgage debt has been originated in the past three
years, thanks primarily to robust home purchase and refinance activity facilitated
by record low mortgage rates. With these loans underwritten on the basis of
recent high collateral values, a decline in home values in some markets could
lead to default activity and losses to residential lenders. In particular,
high-risk borrowers may default in increasing numbers should interest rates
rise and home prices fall. The number of speculative homebuilders also may
elevate construction lending risk in some markets.
Mortgage Debt and Consumer Credit Present Risks
During and after the 2001 recession, persistently low interest rates prompted
households to add consumer debt rather than deleverage. During the year ending
September 2003, U.S. households added nearly $925 billion in debt to their
balance sheets, an increase of more than 11 percent.9 Households
have not assumed debt so quickly since the late 1980s. The sustained increase
in the aggregate financial obligations of homeowners is attributable to both
higher leverage and a higher rate of homeownership, especially among borrowers
with fewer financial resources. The amount of household credit debt is not
worrisome in itself, but its concentration among high-credit-risk households
may pose additional risk to residential lenders. Nonmortgage consumer lenders
also may bear some risk if liquidity issues result in a reprioritization of
debt repayment. Households may make repaying mortgages a higher priority than
repaying unsecured consumer credit, such as credit cards.
Over the past decade, changes in lending standards and the introduction of subprime and high loan-to-value (HLTV) mortgages have allowed new homeowners to qualify for mortgages despite higher overall debt levels and lower down payments. Between 1993 and 2001, the subprime share of all home purchase mortgage originations in metropolitan areas climbed from 1.3 percent to 6.5 percent, while the subprime share of refinance loans jumped from 2.1 percent to 10.1 percent (see Chart 3).10
The popularity of subprime loans carries risk, as subprime loans historically have exhibited default rates on the order of ten times greater than prime loans extended to borrowers with solid credit records.11 HLTV mortgages also have shown higher default rates. In 2002, loans exceeding 80 percent of the home purchase price accounted for more than 35 percent of all purchase mortgages underwritten in 40 percent of the nation's metropolitan areas. In some cities, loans exceeding 80 percent of the home purchase price accounted for over 50 percent of originations during 2002 (see Chart 4).12
Higher interest rates also could have implications for mortgage credit risk. According to the Mortgage Bankers Association, during 2003, 20 percent of conventional mortgages were originated using adjustable rate mortgages (ARMs), despite a historically low fixed-rate environment. After equalizing for differences in the credit quality of borrowers, ARMs offer lower rates than fixed-rate loans during any part of the interest-rate cycle. Thus, they improve affordability and allow buyers to purchase more expensive homes for any given monthly payment. However, these loans also expose homeowners to rising interest rates, since the debt service cost on ARMs is tied to short-term interest rate movements. This feature can increase credit risks for lenders when interest rates and monthly mortgage payments rise.
Although the 20 percent figure cited previously seems modest, it masks a rising trend during 2003. The share of ARMs rose from roughly 14 percent of all conventional mortgages underwritten in January to 32 percent by December, even though conventional 30-year fixed mortgage rates fell during the year. This trend suggests that at least some homebuyers were stretching to keep their monthly payments manageable in the face of rising home prices. Affordability is a persistent issue in the highest-priced U.S. housing markets, such as San Francisco, San Diego, Los Angeles, New York, Boston, Seattle, and Denver. As a result, borrowers in these markets use ARMs more frequently than borrowers elsewhere (see Chart 5).13 Not only do high home prices require more borrowers to seek ARMs, but many of these cities historically have posted some of the widest home price swings (see Map 1). As a result, homeowners in these markets are potentially exposed to both rising monthly mortgage payments and falling house prices if interest rates rise.
Subprime borrowers are another group of homeowners with a disproportionate
exposure to rising interest rates. Based on our estimates, subprime mortgage
borrowers seemed about twice as likely to hold ARMs as conventional borrowers
The exposure of these households to rising interest rates compounds the credit
risk associated with subprime consumer portfolios.
Home equity lines of credit (HELOCs) may carry even greater credit risk than ARMs when interest rates rise or home prices decline. HELOCs, set up as lines of credit from which homeowners can draw up to a maximum loan amount based on the equity in the home, may extend homeowner leverage well beyond 100 percent. In addition, with a variable interest rate that fluctuates over the life of the loan, HELOCs involve greater interest rate risks for homeowners. In contrast to primary lien ARMs, most HELOCs do not have an adjustment cap to limit the size of any payment increase. Furthermore, changes in the prime rate affect the rates of HELOCs immediately.15 With HELOC debt currently at $254 billion and representing a commitment tied to home equity, credit quality in this segment could decline when interest rates rise.16
Although private mortgage insurance (PMI) mitigates the risk of collateral losses to lenders, it does not cover all risks, specifically those from "piggyback" loans. By convention, borrowers are required to supply at least 20 percent down to avoid paying for PMI. But certain loans, typically called 80-10-10 loans, are structured to avoid paying for PMI, and thus no insurance is obtained. Under an 80-10-10, a homebuyer with a 10 percent down payment obtains a loan for 80 percent of the home's purchase price at a standard interest rate and then gets a second, or piggyback, loan at 10 percent of the purchase price, but at a higher interest rate. This type of financing adds leverage over the traditional 20 percent down payment loan at the same time it avoids PMI, a safeguard for lenders. While lending programs such as piggybacks, subprime mortgages, and ARMs have allowed greater opportunities for homeownership, they also may present increased credit risks to lenders, particularly should interest rates rise or home prices fall.
Residential Construction Lending Is an Additional Concern
Another key component of credit risk associated with residential real estate is C&D lending. High levels of speculative residential construction would be of particular concern in a softening housing market, as residential developers and their lenders could face losses when actual sale prices differed from values assumed at the time loans were made. In addition, a recent Federal Deposit Insurance Corporation report17 noted that thinly capitalized builders and newer institutions seeking to expand market share also represent concerns in residential construction lending.
In summary, because home prices have appreciated briskly over the past several years and outpaced income growth, concerns have been voiced about the possibility of a nationwide home price bubble. However, it is unlikely that home prices are poised to plunge nationwide, even when mortgage rates rise. Housing markets by nature are local, and significant price declines historically have been observed only in markets experiencing serious economic distress. Furthermore, housing markets have characteristics not inherent in other assets that temper speculative tendencies and generally mitigate against price collapse. Because most of the factors affecting home prices are local in nature, it is highly unlikely that home prices would decline simultaneously and uniformly in different cities as a result of some shift such as a rise in interest rates.
The greater risk to insured institutions is the potential for increased credit delinquencies and losses among highly leveraged, subprime, and ARM borrowers. These high-risk segments of mortgage lending may drive overall mortgage loss rates higher if home prices decline or interest rates rise. Credit losses may, in turn, spill over to nonmortgage consumer credit products if households prioritize debt repayment to give preference to mortgage payment. Residential construction lending in markets where there is significant speculative building, as well as an abundance of thinly capitalized builders, also may be of concern, especially when the current housing boom inevitably cools.
Cynthia Angell, Senior Financial Economist
The FDIC's Economic and Banking Outlook in Charts
The U.S. economic expansion accelerated and became more balanced in 2003. Consumer spending gains continued apace, while business investment rebounded, commercial real estate investment was less of a drag, and the weaker dollar helped to make international trade a positive contributor to growth during three of four quarters last year. But job growth remains the most obvious missing link to a sustainable, vigorous economic expansion. During the five months ended January 2004, nonfarm payroll growth averaged only 73,000 per month. Gains at more than double that pace are needed to bring unemployment down appreciably. The factory sector, which is starting to show signs of improvement in orders and output, has been the main drag on overall job growth since the recession ended in November 2001.
Looking forward, real GDP is expected to gain about 4.5 percent this year on stronger growth in business investment, inventory building, and exports. Consumer spending should maintain a moderate pace of growth and may be boosted in the short term by a favorable tax refund season. Most analysts expect inflation and interest rates to remain relatively low. The major risk to this favorable outlook remains the employment situation. Without stronger job growth, consumer income, confidence, and spending could weaken, dampening the prospects for economic growth beyond midyear.
During the first nine months of 2003, more than 94 percent of FDIC-insured institutions were profitable, and the industry's return on assets (ROA) reached a record 1.37 percent. The FDIC's outlook for banking in 2004 remains positive, given favorable economic trends, improving asset quality, strong earnings, and high capital levels. However, the industry does face some challenges in the coming year.
High and increasing levels of long-term assets have created some exposure
to rising interest rates. Moreover, there is a question regarding whether the
household sectorthe engine for banking growth through the 2001 recession and
beyondcould falter. Sources of concern about households include rising indebtedness,
the near-record pace of personal bankruptcy filings, and the possibility that
consumer spending could tail off once the effects of the 2003 tax cut and the
2003 mortgage refinancing boom run their course.
Another challenge for the banking industry is in commercial real estate, where lending concentrations are high and market fundamentals have been poor. Offsetting these challenges is a solid financial picture for the industry, where measures of earnings and capital are generally at or near record levels.
Norman Williams, Chief, Economic Analysis Section
Allen Puwalski, Chief, Financial Institutions Analysis Section
Jack Phelps, CFA, Regional Manager
Scott Hughes, Regional Economist
1 Sales of existing homes rose to record levels in 2003, surpassing previous records set in 2002. Sales of new single-family homes topped 1 million units, establishing a new record high for the third consecutive year. In addition, according to the National Association of Home Builders, new home starts rose in 2003 to 1.85 million, the highest number of housing starts since 1978.
2 Karl E. Case and Robert J. Shiller, Is There a Bubble in the Housing Market? An Analysis, Washington, DC: Brookings Institution, 2003.
3 Federal Reserve Board, Flow of Funds data.
4 Stephen Malpezzi, "Housing Prices, Externalities, and Regulation in U.S. Metropolitan Areas," Journal of Housing Research, Vol. 7, Issue 2, 1996, pp. 209-241.
5 Gary V. Engelhardt, Nominal Loss Aversion, Housing Equity Constraints, and Household Mobility: Evidence from the United States, CPR Working Paper Series # 42, Syracuse, NY: Center for Policy Research, Syracuse University, 2001.
6 D. Genesove and C.J. Mayer, "Nominal Loss Aversion and Seller Behavior: Evidence from the Housing Markets," Quarterly Journal of Economics, 116, 2001, pp. 1233-1260.
7 Mortgage Bankers Association/Haver Analytics.
8 FDIC Call Report data, September 30, 2003.
9 Susan Burhouse, FYI: Evaluating the Consumer Lending Revolution, Washington, DC: Federal Deposit Insurance Corporation, September 17, 2003.
10 Joint Center for Housing Studies of Harvard University, The State of the Nation's Housing, Cambridge, MA, 2003.
12 Federal Housing Finance Board. Monthly Interest Rate Survey, 2002.
13 Allen Puwalski and Norman Williams, FYI: Economic Conditions and Emerging Risks in Banking, Washington, DC: Federal Deposit Insurance Corporation, November 4, 2003.
14 Our estimates are based on the following: As
of the third quarter 2003, 56 percent of the total mortgage loans (on a dollar
LoanPerformance subprime database were "non-fixed" (ARMs and hybrids).
Given that over three-quarters of all mortgage loans were underwritten
in the past three years, we used the share of conventional ARMs
underwritten (data from the Mortgage Bankers Association weekly
survey, on a dollar basis) during those years as a rough approximation
for the total outstanding share of conventional mortgages in ARMs.
These MBA data indicate that ARMs accounted for roughly one-fourth of
the value of conventional mortgages underwritten between 2001 and
third quarter 2003, or roughly one-half the LoanPerformance subprime
15 Certain HELOCs have guaranteed fixed introductory rates, but these rates typically are binding for only a few months.
16 The amount of debt tied to HELOCs, as measured by revolving home equity loans at domestically chartered commercial banks, jumped from $106 billion in the first quarter of 2000 to $254 billion in the third quarter of 2003, averaging an almost 35 percent annual increase. Source: Federal Reserve Board.
17 Allen Puwalski and Norman Williams, FYI: Economic Conditions and Emerging Risks in Banking, Washington, DC: Federal Deposit Insurance Corporation, November 4, 2003.