Since the collapse of the housing bubble and onset of the Great Recession, residential foreclosures in the US have been at levels not seen since the Great Depression. Given the great harm that these foreclosures have inflicted on the individuals and communities affected it is of great importance to have a clear understanding of the relative importance of the many factors that contributed to the loss of these homes to inform efforts by the public and private sectors to better prepare and support homebuyers in the future. In fact, there has been a great deal of analysis of mortgage outcomes over the last several years that have identified how loan terms, lending channels, and borrower characteristics at origination have been associated with greater risk of default and foreclosure (see, for example, Mian and Sufi, 2009, Demyanyk and Van Hemert, 2011, and Ding, Quercia and Ratcliffe, 2011). But these studies cannot provide a full accounting of the factors contributing to home losses as available data do not have access to complete of borrower circumstances and how they changed over time, including, for example, whether the borrower has experienced a loss of income, an increase in other financial obligations, or a change in household composition due to death or divorce. Nor do these studies have access to a complete accounting of the household balance sheet to determine whether the household has additional financial resources that could potentially be tapped to fend off loss of the home. Finally, analysis of loan outcomes ignores forced exits from homeownership that do not entail a foreclosure.

The purpose of this study is to provide a more complete assessment of the factors contributing to whether owners were able to sustain homeownership during the boom and bust years through analysis of panel data on households that provides rich data on financial circumstances. Specifically, the study will use the PSID for the period 1999-2011 to estimate survival models for homeownership spells to identify the relative importance of different factors that are likely to have triggered the exit. The analysis will build on a number of previous studies that looked at similar questions for earlier time periods, including Reid (2004), Haurin and Rosenthal (2004), Turner and Smith (2009), and Boehm and Schlottmann (2014).  But none of the existing studies have examined this latest period. An important emphasis of the present study will be to explore two factors driving exits from homeownership that were less prominent important in earlier periods—namely, the growing prevalence of mortgages that permitted housing payments that were relatively unaffordable at origination and the very high prevalence of underwater borrowers in the wake of the substantial decline in home prices.  At the same time it will also explore the relative importance of unemployment and income loss, which have often been downplayed as key drivers of the foreclosure crisis. The study will also have an explicit focus on assessing how factors associated with exits from homeownership differ by race/ethnicity and income.