Deconstructing a Mortgage Meltdown
Abstract
Foreclosure rates in the United States have risen steadily for decades, reflecting both the long-run loosening of mortgage underwriting and the shifting economic fortunes of American communities. This paper develops a decomposition that separates these two forces by expressing foreclosure rates as the product of local economic conditions and underwriting quality. Using national servicing data and forward-looking regional risk indices, we show that underwriting deteriorated in two distinct waves: first in the mid-1990s, when new credit-scoring technologies expanded access to high-LTV lending, and again after 2002, when securitization enabled a rapid expansion of nontraditional loans. The first deterioration was masked by unusually strong economic conditions; the second was hidden in unobservable underwriting variables that investors could not monitor. When house prices fell, both forces reversed simultaneously, producing the unprecedented surge in defaults after 2005. The results highlight how technological optimism, misaligned incentives, and local economic shocks interacted to generate the mortgage meltdown, and they underscore the need for institutions that discipline underwriting even when recent performance appears benign.
The modern American mortgage market has been shaped by two powerful and often conflicting forces: technological innovation in underwriting and the shifting economic fortunes of the places where borrowers live. Over the past several decades, advances in credit scoring and automated underwriting made it easier for lenders to expand credit. These tools lowered the cost of evaluating borrowers, but they also encouraged a steady relaxation of lending standards. Foreclosure rates rose accordingly.
Understanding the mortgage meltdown requires separating the role of underwriting from the role of economic conditions. Borrowers default when they face both financial stress and insufficient home equity. A borrower with stable income and rising home prices rarely defaults; a borrower with declining income and falling collateral values often does. Our goal is to disentangle these forces by decomposing foreclosure rates into two multiplicative components: one capturing the economic environment facing borrowers, and the other capturing the quality of underwriting embedded in the loans.
The Long-Run Rise in Foreclosures
Foreclosures have been rising for nearly three decades. From 1979 to 2002, the share of loans entering foreclosure quadrupled. This long-run trend reflects more than just business cycles. It reflects a structural shift in the mortgage market: lenders gradually expanded credit to riskier borrowers, and underwriting standards loosened accordingly.
Yet the timing of foreclosure spikes reveals the powerful role of economic shocks.
Foreclosures rise above trend during recessions-1981-82, 1990-91, 2001-02-and fall
below trend during expansions. The surge after 2006 stands out because it dwarfs earlier
cycles. The magnitude of the increase suggests that the Great Recession was not just
another downturn but a shock that exposed deep vulnerabilities in the mortgage system.
The deterioration in mortgage performance was not confined to subprime loans. Subprime foreclosures rose earlier and more sharply, but prime loans followed the same pattern with a short lag. This parallel movement indicates that both markets were responding to a common set of forces-falling home prices, rising unemployment, and the unwinding of a decade-long boom. Subprime borrowers were simply the first to feel the stress, much like the canary in the coal mine.
The Central Role of House Prices
Modern mortgage theory treats the borrower as holding a put option on the home. When
home prices rise, borrowers in distress can sell rather than default. When prices fall,
especially below the value of the mortgage, default becomes the rational choice.
For most of the period from 1975 to 1997, real house prices moved within a narrow band. But from 1997 to 2007, real prices rose more than 40 percent above their long-run level. This extraordinary boom reduced defaults and lulled lenders into believing that their underwriting models were more robust than they were. Models calibrated on a decade of
rising prices inevitably underestimate risk.
The acceleration in house prices after 1999 was particularly dramatic. In some
metropolitan areas-San Diego, Miami-prices soared and then collapsed. In others¬
Detroit-prices barely rose at all, but unemployment surged, producing a different but equally powerful form of financial stress. These local differences are essential for understanding the geography of the foreclosure crisis.
Measuring Local Economic Stress
To separate the effects of underwriting from the effects of economic conditions, we rely on UFA's Default Risk Index, which estimates how a constant-quality loan would perform under the economic conditions prevailing in each state and year. The index varies widely¬from 60 to 270-implying that economic conditions alone can cause more than a fourfold difference in expected defaults.
Two patterns stand out. First, economic conditions improved steadily from 1990 to 2002, creating a benign environment that masked the deterioration in underwriting. Rising home prices and strong labor markets allowed even risky loans to perform well. Second, after 2002, the index rose sharply as house prices flattened and then fell. The index anticipated the downturn because it incorporates forward-looking house-price forecasts. This makes clear that the economic environment was not merely a backdrop to the crisis-it was a central driver.
A Framework for Decomposing Foreclosures
Foreclosures occur when borrowers face both financial stress and insufficient home equity. Our model captures this interaction by treating the foreclosure rate as the product of three components: the baseline hazard, the economic environment, and the underwriting characteristics. Because the Mortgage Bankers Association reports foreclosures only in aggregate-across vintages and borrowers-we infer underwriting quality indirectly by conditioning observed foreclosure rates on UFA's economic indices. The remaining variation reflects underwriting.
This approach mirrors the logic of urban economics: when we cannot observe individual behavior directly, we infer it from aggregate patterns and the incentives that shape them.
What the Model Reveals
The results are striking. When we regress foreclosure rates on the economic indices and extract the year fixed effects, we obtain a time series of underwriting quality. This underwriting index shows two major periods of deterioration.
The first occurred in the mid-1990s. During this period, lenders expanded access to low-down-payment loans and embraced credit-scoring technologies. Observable underwriting variables worsened, but the effects were masked by strong economic conditions. Rising home prices and falling unemployment made even risky loans appear safe.
The second deterioration occurred after 2002. This time, observable underwriting variables did not change much. Credit scores and loan-to-value ratios remained stable. Yet defaults rose sharply. The only plausible explanation is that underwriting deteriorated along unobservable dimensions-documentation quality, income verification, appraisal standards-precisely the areas where securitization created moral hazard.
The decomposition shows that by 2004, underwriting erosion alone would have doubled foreclosure rates relative to 1990. But because economic conditions were so favorable, only a modest increase was observed. After 2005, both underwriting and economic conditions deteriorated simultaneously, producing the unprecedented surge in defaults.
Prime vs. Subprime: Different Levels, Same Mechanisms
The model also reveals that prime and subprime loans respond similarly to economic shocks, though at different levels. Subprime borrowers default earlier and more frequently, but the elasticity of defaults with respect to economic conditions is actually higher for prime loans. This suggests that prime borrowers, who typically have more equity and more stable incomes, are more sensitive to changes in collateral values and employment conditions.
The parallel movement of prime and subprime defaults reinforces the idea that the crisis
was not confined to one segment of the market. It was a systemic event driven by the
interaction of deteriorating underwriting and worsening economic conditions.
A Structural Break Consistent with Moral Hazard
Before 2002, underwriting tended to tighten when economic conditions weakened. This
countercyclical pattern is consistent with lenders learning from recent performance. After
2002, this pattern broke down. Underwriting quality continued to erode even as economic
conditions worsened. This shift is consistent with a structural break driven by
securitization.
When originators can sell loans quickly into complex securities, they bear little of the long-term risk. Investors, who ultimately bear the risk, cannot observe the full set of underwriting variables. This separation of knowledge and incentives creates classic moral hazard. The rapid growth of nontraditional lending after 2003-subprime, Alt-A, home equity-magnified this problem. Originators faced strong incentives to expand volume, while investors struggled to monitor quality.
As long as house prices rose, this erosion remained hidden. When prices fell, it became
catastrophic.
The Broader Lessons for Mortgage Markets
The crisis reveals how deeply mortgage markets are shaped by the interaction of
technology, incentives, and local economic conditions.
• Technology made it easier to expand credit but encouraged overconfidence when calibrated on unusually favorable periods.
• Incentives in securitization separated originators from risk, creating moral hazard.
• Local economic conditions masked fragility duringthe boom and exposed it during the bust.
The decomposition developed here provides a way to measure these forces separately. It shows that the mortgage meltdown was not the result of a single misjudgment but the predictable outcome of a system in which optimism was rewarded, information was incomplete, and economic conditions were unusually favorable for an unusually long time.
The crisis was a reminder that markets are most vulnerable when they appear most stable. When rising home prices make every loan look safe, underwriting standards erode. When models are calibrated on good times, they underestimate risk. When originators do not bear the consequences of their decisions, quality declines. And when economic conditions reverse, the consequences are severe.
The challenge for policymakers and investors is to design institutions that recognize this
reality-institutions that reward accurate risk assessment, discourage moral hazard, and
incorporate the full range of economic forces that shape borrower behavior.
Based on ANDERSON, C.D., CAPOZZA, D.R. and VAN ORDER, R. (2011), Deconstructing a Mortgage Meltdown: A Methodology for Decomposing Underwriting Quality. Journal of Money, Credit and Banking, 43: 609-631. https://doi.org/10.1111/j.1538-4616.2011.00389.x