The U.S. housing bubble that ended in early 2007, characterized by abnormally robust growth in home prices, was notable for showing what can happen when housing market investors take on too much risk. Granted, risk is a permanent fixture in the investing landscape. Indeed, it can rarely (if ever) be avoided. But the bubble, and the financial crisis that unfolded in its wake, exposed the perils of excessive risk exposure, for which financial markets ultimately paid a heavy price. The sheer heft of the risky behavior leading to the crisis is well understood, particularly in mortgage markets, where underwriting standards loosened generously.1 Less understood, however, is that underwriters were not alone in expanding their risk appetite; some borrowers did so as well, brazenly misrepresenting their intended occupancy status when filling out mortgage applications. In “Owner-Occupancy Fraud and Mortgage Performance,” the Philadelphia Fed’s Ronel Elul and his coauthors investigate these misrepresentations, showing how a significant number of mortgage applicants identified themselves as owner-occupants (that is, people who intend to live in the homes they own) despite their apparent status as investors. Further, their analysis shows that such misrepresentations persisted in the years following the housing bubble, with more recent mortgage originations revealing the same troubling behavior: Rather than live in the homes being purchased, dishonest borrowers treated them solely as investments. In doing so, they committed occupancy fraud, and they continued to do so even in the years after the housing bubble burst.
To identify fraudulent mortgage borrowers, Elul and his coauthors focus on isolating those who 1) are purchasing homes; 2) claim to be owner-occupants; 3) have multiple first liens; and 4) do not move to a new address in the year after their mortgage is originated. The logic behind this methodology is that, while these mortgage borrowers claim to be owner-occupants and, therefore, should be moving into their new home, it seems that they continue to live in their original residence. In having multiple mortgages and not moving, those classified as fraudulent more closely resemble those who declare themselves as investors at the outset.
Elul and his coauthors show that including these dishonest borrowers, who made up over 7 percent of all homebuyers at the peak of the bubble in early 2006, effectively doubles the size of the investor population. They also show that a substantial amount of fraud is found across all major mortgage issuance channels — whether they be private or backed by government-sponsored entities. Although they find that fraudulent investors were especially active during the housing bubble, they also show that fraud persisted through the end of their sample (December 2017).
After identifying fraudulent investors, the authors try to understand their incentive for committing fraud. They suggest that their incentive was to obtain a mortgage on better terms than what was available to investors. As a result, fraudulent investors have larger mortgages, lower credit scores, higher loan-to-value ratios, and lower interest rates than borrowers who declare themselves to be investors.
This confluence of negative factors created a substantial amount of risk for mortgage markets to bear, and this risk was exemplified in several ways. First, fraudulent investors defaulted on their loans more often than other types of borrowers, even after taking into account that they had riskier loans. The authors also write that “their default decisions were less likely to be driven by an inability to pay…” They demonstrate this by first showing that fraudulent investors were less likely to draw on other sources of credit, such as bank cards, to try to avoid default. Instead, the authors write, they were more likely to be sensitive to a decline in the value of their homes than other types of borrowers, who were more willing to hold onto their homes even following a large decline in value. In this sense, they defaulted more “strategically” than other borrowers.
Could other, less nefarious triggers contribute to a correlation between fraud and default? With this question in mind, Elul and his coauthors consider alternative explanations for their primary findings. They look, for instance, at the possibility of “accidental fraud,” in which mortgage applicants do not set out to commit fraud but nonetheless find themselves financially overstretched — perhaps from being unable to sell their original homes. They test for this by considering various measures of housing market weakness and show that they “have limited explanatory power for fraud.” After running additional experiments in search of other alternative explanations, the authors confirm the lack of any substantial evidence that would overturn their key results.
Throughout their analysis, the authors produce consistent evidence of elevated risk levels in mortgage markets, especially during boom-and-bust cycles in housing prices. (Not inconsequentially, mortgage fraud was found to be notably high in parts of the U.S. that witnessed particularly aggressive housing bubbles, such as certain markets in Florida and Arizona.) “We demonstrate,” the authors explain, “that occupancy fraud is much more widespread and persistent than shown in the previous literature.” Given the prevalence of fraud, they suggest that researchers pursue further studies, particularly to consider why borrowers choose a fraudulent path in the first place. “One area for future research,” the authors write, is “whether behavioral characteristics may help explain why some borrowers were led to speculate in housing markets through fraud.”
In their work, Elul and his coauthors take a broad, inclusive look at mortgage markets, analyzing hundreds of thousands of loans to document the prevalence of occupancy fraud among mortgage applicants. Their study adds depth to the existing knowledge of risk drivers in mortgage markets, opening the door to a better understanding of risk's root causes. Although they focus on mortgages originated in the past, the authors stress that their findings have important ramifications for current mortgage markets as well. “We show,” they note, “that fraud continues to remain a concern, appearing even in recent originations.”
- The views expressed here are solely those of the authors and do not necessarily reflect the views of the Federal Reserve Bank of Philadelphia or the Federal Reserve System.
- For a comprehensive risk assessment, see the 2011 Financial Crisis Inquiry Report released by the Financial Crisis Inquiry Commission.