Painful Debt

Homeowners borrowed 25 cents for every dollar increase in house prices. That adds up to $1.25 trillion in household debt—a very large figure given that previous research has generally concluded that households do not borrow heavily because of an increase in their home equity.

But not all types of consumers responded in the same way, say the authors. Homeowners with the lowest credit scores and those who relied heavily on credit cards were the most eager borrowers. These overstretched households may have had trouble obtaining credit before the housing boom, so they were happy to use the value gained by their homes as collateral for fresh loans. Another, perhaps more worrisome, interpretation is that these homeowners had self-control problems; the lure of easy credit was simply too tempting. The result of this borrowing binge was that home-equity loan defaults were responsible for almost 40 percent of all new defaults in 2006 to 2008.

Finding an Effect of Home Value on Borrowing

It is not obvious that homeowners should feel wealthier when their home value increases. This is because homeowners are also consumers of housing goods. An increase in their home value implies a higher cost of housing services going forward, which cancels out the new wealth they have in their own home. In this case, higher house prices do not give homeowners an incentive to borrow against a larger home equity.

However, homeowners who are close to retirement and have planned all along to consume at least part of their housing capital before they die are more likely to take advantage of a home equity loan when house values suddenly appreciate. The same is true for individuals with difficulties obtaining credit; these homeowners want to borrow more but are unable to do so because of limited collateral.

Looking at the growth in household debt of a random sample of 74,000 homeowners from 1997 to 2008 and house prices during this period, it is clear that both figures moved in the same direction. While this trend suggests that homeowners may indeed withdraw equity when house values go up, an increase in expected earnings also can drive changes in both house prices and borrowing. For example, a homeowner anticipating a higher income in the future may want a bigger house today, but can only buy that house by borrowing more. As a result, if many people become more optimistic about their future income, house prices and household debt can both rise without the increase in house prices causing the increase in debt.

To assess the extent to which home prices directly affect household borrowing, Mian and Sufi used housing supply data to control for demand-side factors that affect both home prices and borrowing. The intuition is that given a certain demand for housing, prices will rise much faster in areas where housing supply is "inelastic," i.e., locations where houses cannot easily be built to meet demand, perhaps because land is scarce. They ranked cities based on the elasticity of housing supply, and found that in places where housing supply is very inelastic (the bottom quartile), house prices from 2001 to 2006 more than doubled, while prices in areas with an abundant supply of homes (the top quartile) did not see much inflation.

Homeowners in cities where house prices rose quickly borrowed more, but the response varied depending on the type of consumer. Households with the lowest credit scores and who were heavy credit card users, as measured by the fraction of available credit used, responded more to the fast-paced rise in home values. In contrast, there was almost no effect on home-equity borrowing for the homeowners who carried very low balances on their credit cards. Their borrowing behavior was similar to homeowners who lived in cities that experienced almost no increase in house prices.

Credit scores determine the availability and pricing of consumer credit, and consumers below critical thresholds are often unable to obtain financing at reasonable interest rates. Thus, a low credit score is a typical characteristic of credit-constrained individuals, as is carrying a credit card balance close to its limit. The sudden increase in house prices gave these consumers the collateral they needed. "Before the housing boom, these homeowners could not borrow because creditors would not give them loans that were unsecured," says Sufi.

However, low credit scores and heavy credit card use also fit the description of individuals who are prone to borrowing too much, which most likely will lead to future defaults. Such homeowners borrow aggressively against an increase in housing wealth to finance current spending. Whether these characteristics reflect traditional credit constraints or simply a lack of self-control is an open question for future research, say Mian and Sufi. Additionally, younger homeowners were keener to borrow when house price inflation accelerated. This is inconsistent with the life-cycle theory of consumption, which predicts that younger people save rather than borrow to accumulate wealth for retirement. In the absence of liquidity constraints, older homeowners should be more willing to extract cash from the increased home values.

The Pain in Housing Gains

Mian and Sufi estimate that homeowners borrowed more than a trillion dollars from 2002 to 2006 through home equity loans, and that homeowners did not use the extracted funds to pay down other debts or purchase financial assets. This suggests that loans were used primarily for spending, which is supported by other studies indicating that households use home equity loans mainly for home improvements and other consumer items.

As the housing market started to unravel in 2006 to 2008, homeowners who had piled on debt began defaulting on their loans. In particular, the study finds that the default rate for low credit-quality borrowers in cities where house prices increased the most jumped by 12 percentage points, compared with only a 4 percentage point increase in areas that experienced little price appreciation.

High rates of default suggest overborrowing, and as Mian and Sufi point out in another paper, households burdened by debt is an important impediment to growth. In fact, they find that changes in household leverage at the county level served as an early and powerful predictor of the onset and severity of the recent recession.

Counties with the highest and lowest household debt-toincome ratios (the top and bottom ten percent, respectively) both experienced a sharp drop in auto sales in the worst part of the recent recession. Auto sales in counties with the highest debt-to-income ratios remained weak in 2009, but sales in counties with the lowest debt ratios experienced a robust recovery. Similarly, residential investment in counties with the highest household debt-to-income ratios remains well below pre-recession levels, while counties with the lowest debt ratios almost completely avoided a decline in residential investment. Rebounds in residential investment and durable consumption are important because they often pave the way to economic recovery.

"We believe that the most important factor for understanding the severity of the economic downturn of 2007 until today is high household debt levels," says Sufi. As the authors show, withdrawing home equity can explain a very large fraction of the increase in household debt and the ensuing default crisis.