The spike in mortgage defaults during and subsequent to the financial crisis had a meaningful impact on the real economy, both directly and indirectly. The direct effects are more easily measured; millions of defaulting borrowers lost their homes and their equity. Indirect effects are more nuanced, but the indirect effects that can be measured indicate that foreclosures negatively impact the value of neighboring homes.
While the consequences of mortgage defaults are observable, understanding the cause of defaults has been less clear. Every mortgage default can be attributed to one of two explanations. The first is that borrowers exercise their option to strategically default—defined as borrowers making the choice to stop paying their mortgages despite having the capacity to do so. The second explanation is that borrowers are unable to pay their mortgages due to a liquidity shock (e.g. loss of job, medical emergency, unforeseen financial shock, etc.). The question analysts, policymakers, and academics have been grappling with since the Great Recession is to understand what fraction of mortgage defaults are strategic defaults and what fraction are the result of borrower liquidity shocks. And if some defaults are purely strategic, how could an analyst or model detect a strategic defaulter ex ante?
Due to data limitations, the evidence on the fraction of defaults that are strategic versus liquidity-driven has been mixed until, in my opinion, just recently. Peter Ganong and Pascal Noel, two up-and-coming scholars at the Booth School of Business at the University of Chicago are putting the finishing touches on a new study that addresses the strategic vs liquidity shock question head on. The study amasses administrative data from JP Morgan Chase that combines borrowers’ checking and deposit account data with loan servicing records for each borrower. These combined data allow for careful analysis of the patterns of borrowers’ incomes throughout the life of a mortgage. What does the study show? In almost every instance of mortgage default, defaulting borrowers experienced a substantial shock to their incomes in the six- to nine-month period prior to default. This isn’t to say that every borrower that had an income shock defaulted—some borrowers are able to dig themselves out of a financial hole. Rather, their study suggests that it is very rare for defaults to exist without a preceding income shock.
This result is in contrast to several previous studies that argued for the existence of strategic mortgage defaults. The most relevant comparable study published in the Journal of Finance in 2017 was written by a team of authors from the Federal Reserve Board and Goldman, Sachs & Co (Neil Bhutta, Jane Dokko, and Hui Shan). These scholars document that borrowers with CLTV ratios in excess of 150 percent default strategically. While theoretical models predict that borrowers should strategically default at much lower CLTV levels, the Bhutta, Dokko, and Shan study still documented the existence of strategic defaults.
One explanation for the difference between the University of Chicago study and the Fed study is the composition of borrowers. Loans from the Fed study were all subprime, as compared to the Chicago study that analyzed a larger fraction of prime borrowers. However, despite differences in sample composition, the novelty of the Chicago study comes from their ability to analyze a time series of borrower income data prior to defaults, while the Fed study does not. The lack of income data makes it difficult for the Fed study to rule out the possibility that defaulting borrowers, while undoubtedly extremely underwater on their mortgages, may still have experienced substantial negative income shocks prior to defaulting.
What are the implications of the new study? When possible, analysts should closely monitor the stability of borrower income. Incomes will paint a much more relevant, real-time picture of borrower health, as opposed to slower-moving credit scores or sometimes difficult-to-interpret macroeconomic conditions. The study also implies that because of the stigma and social costs of bankruptcy, borrowers are less likely to strategically default than models might predict. Summing up, the latest and best research on this topic seems to suggest that as long as borrowers can make payments, they do.
Taylor Nadauld, PhD
Chief Economist, Visible Equity